Part 2: Did I MIS-DIAGNOSE This Mercedes?? (GL450 no A/C - Crazy REPAIR!)

The owner brought his 2012 Mercedes Benz GL450 to my shop along with the original A/C compressor and junkyard replacement HVAC module.
Let's get out the big guns (PicoScope!), and take a DEEP DIVE to see why the HVAC controller is NOT HAPPY with the A/C compressor, throwing a SHORT CIRCUIT code on the control wire.
Can we get this A/C system to function PROPERLY with the parts on hand??
IVAN'S PICO WAVEFORMS:
drive.google.c...
THINKTOOL PROS: ***$170 OFF!!***
www.amazon.com...
TS100 SOLDERING IRON:
www.amazon.com...
Enjoy!
Ivan

[@TheRealJerseyJoe]

Picoscope + Radio Shack + Russian ingenuity = Success ! Well done Ivan, great fix :-)

[@satamanschmidt3428]

How many mechanics on Earth would have thought of this? You're one in a million Ivan.

[@PineHollowAutoDiagnostics]

This was more of an electrical engineering problem 😉

[@rj.parker]

Good mechanics know 12v inductors need reverse biased suppression diodes. That includes relays. These days you can buy a 12v relay to add on something or you can buy an auto 12v relay that has the diode built in. Ivan knows electrical at close to an EE level and is a degreed engineer, mechanical I believe.

[@rj.parker]

Most factory circuits have the suppression components built into the controller. Generally so effective that some cars like Toyota can take reversed 12v connections without bowing their ecus.

[@1leggedmedia]

I am a truck mechanic and refer mechanic disabled since 2007 and I wouldn't have figured this out like that. He is a better mechanic then I am and I did it for 30 years and specialized in refer units

[@Michigan_Tactical]

I was thinking of it but I'm not a mechanic. I'm electronic tech.

[@Michigan_Tactical]

For anyone that doesn't know, PIV is peak inverse voltage. It's the maximum reverse voltage the diode will handle before starting to break down. 50piv is just right for this application.

[@John-dp3ln]

And if you exceed the piv at all, it will nuke the diode. This is a very important specification. In this case the diode is essentially across the coil, which is correct, and piv doesn’t come into play. Diode ‘eats’ the arc as soon as the reverse voltage exceeds .6 volts.

[@amundsen575]

PIV is kind of an older term, those radio shack diodes are a bit old! datasheets now use the term peak reverse voltage, Peak Repetitive Reverse Voltage , all the same thing. just like when we used the term kilocycles, now we use kilohertz

[@youutartur1104]

I know for a fact straight from a Nissens specific training seminar that vintage of Mercedes AND Volvo with a solenoid valve compressor comes with and without a flyback diode. Their compressors come with a subharness included that contains a diode. I don't recall where you purchased the compressor or what brand, but that class immediately came to mind. This is why training is important, I only do BMW products however information like that will save your ass one day. Another fun fact, new Volvos with 1234yf have a capacitance sensor in their AC system due to the fact refrigerant flows through the battery packs on their electric cars. Absolutely never use PAG oil (because it is conductive), or you have just bought the shop a shiny new brick that it's start is disabled and needs it's entire AC system replaced.

[@jamesanderson2176]

I'm impressed that the controller survived all those negative spikes!

[@trevorvanbremen4718]

I completely agree!!! Furthermore, I'd be willing to bet that the 'junkyard' unit died for EXACTLY that reason!

[@rkan2]

Automotive electronics are usually build to handle a lot.

[@jcos55chev19]

My thought is that the original compressor has a diode built into it that shorted and then blew open. The replacement compressor doesn't have a diode in the harness, (probably a replacement full time non-PWM compressor). If you remember the original diagnostic code alluded to a shorted circuit. The diagnostic code wasn't 100% correct in this case but close to correct. If the schematics were more accurate, they might have shown either a diode coming into the compressor at the short pigtail or shown the compressor with a diode in it.

[@TheOriginalDaveJ]

This comment certainly answers my questions. Wonder if Ivan went looking for the built in diode after the fact.

[@xalty1200]

benz doesn’t duty cycle these things like the other brands that use a denso vd compressor. the dot on the valve means it has an integrated diode and it lives life at battery voltage.

[@hightttech]

Diode in compressor might have shorted, and there's your original short.

[@10100rsn]

The diode is probably built-in to the mosfet that controls the PWM. The internal diode can fail while the switching side of the mosfet still works. The mosfet in the junkyard/ebay controller was probably worse off and the diode took out some of the mosfet. Would be better to use a Schottky Diode in this application but I'm not sure what type he used.

[@xalty1200]

@@10100rsndiode in-line with the load only benz versions of the denso seu compressors have diodes on their vd valves. aftermarket pos had the wrong valve in it

[@andrewthompsonuk1]

Always watch out for diodes built into relays as well. A relay swap can blow a PCM if no diode is on it.

[@robertlacob8831]

Great job Ivan! At 72, I'm no electrical genius anymore, but I can see common sense and understand things that are stupid, even bad designs! And yeah, ALL BRANDS have and continue to build bad designs after 100 years! Just amazes me!

[@brucel76]

The manufacture used several suppliers for the AC compressor! This solenoid is only compatible with select Denso 6SEU16C, 7SEU16, 7SEU17C compressors equipped WITH A DIODE, held in with a snap ring and are visually identical.

[@johnthrossel9981]

Just went to a nissens (ac component manufacture) training and he said this is common on vw’s. Do your homework before just throwing any ac compressor in your German automobile. Some of the control valves have a diode built in and some don’t.

[@belljoxer]

Ivan, good catch and fix on the back EMF. However for a device that monitors current your test lights lights are problematic. When you first apply power to a incandescent lamp there is very little resistance and this causes excessive current to flow. As soon as the filament heats up the resistance increases and the current flow decreases. This is why you initially saw 4 amps on the scope and then it settled down to 1 amp. The controller likely saw this 4X normal current flow and shut down to protect the controller before the lamp reached it's steady state 1 amp draw.

[@PineHollowAutoDiagnostics]

Correct 👍

[@brucepeebles4939]

As soon as I saw the "Back EMF" spikes, I *knew* there was missing diode. As a general rule, ANYTIME a coil is driven with sold-state device, a snubber-diode to ground is standard electronic-design. Technically, the diode should be AS CLOSE AS POSSIBLE TO THE COIL so the backEMF has no chance to propagate down the wire and crosstalk into other circuits.

[@JD-iu3vi]

A better install of the diode would have been to solder pigtail wires to the diode. Cover diode and solder connections with heat shrink then connect pigtails to the wires. Also, a label attached indicating what the component does.

[@timm7885]

I have ordered my "Ai Ivan" model. Well done, I am at a loss at how this channel is not at 500K subs!

[@Graham_Wideman]

Well, there is already a "Hey, Ivan!" service, called PHAD :-)

[@hugh007]

I once found an open diode in a Chrysler wiring harness. No AC clutch operation. It was well hidden in the harness. The dealer didn't have one and didn't even know it was there. I used a 1N4007 or something as a repair and taped it up again. I also seem to remember a diode installed on some Fords right at the clutch connector. Good job Ivan.

[@donkaiser6704]

Used in that configuration, it's called a "flyback" diode.... always good practice in circuits that drive an inductive load (even relays). There was probably one inside the controller that blew-out at some point in the car's history.

[@JoseSilveira-newhandleforYT]

Good catch, Ivan! The internal diode could have been blown on a careless bench test with the wrong polarity, or just had a premature fail in the silicon. You definitely need an anti-parallel diode on inductive loads controlled by semiconductors - it was sheer luck that running without the diode didn't damage the output transistor in the controller. An oscilloscope (and the brain to interpret it) is fundamental to catch these faults. Great repair!

[@briankrug2614]

I have a gl450 2012 and have been dealing with this issue for 5 years now. It was very intermittent at first, but with each season, it would happen more frequently... This year, just about every trip over 20 min it would happen. I have had it to my local shop, but they are not really Mercedes mechanics and did the best that they could with it. They never threw the parts canon at it, thankfully. I was about to take it to our local Mercedes dealer to look at it, but the last time I had it there, it cost over $3000 for a suspension issue by replacing bags and a air manifold. Never corrected the issue. My local mechanic figured out the leveling sensors were not adjusted correctly. His $25 fix corrected that issue. So... long story short, I was very reluctant to take it to the dealer. Then I found these vids.... your explanation of the issue was spot on. And, your explanation of why was great and made a lot of sense. Who would of thought of inductive back voltage would be the issue. I may be buying an oscilloscope... through, I would probably never use it, as I am not an electronics engineer or an auto mechanic. Just know enough to get myself in trouble. Regarding the oscilliscope, what are you using. My foxwell reader was showing the same short or open circuit code on the compressor, and the Request To Component was initiating, would control for a bit, then fall to 0. Total cost to fix, $300 for the Foxwell OBD reader (which I actually purchased for the suspension issue) + $0.12 for the rectifier diode (really $6.00 because I had to buy 50 of them). Thank you for taking the time to share your knowlege... It saved me (and I am sure many others) a ton of money!!!

[@vpimike2646]

Oh yeah, I forgot to mention that my hunch was right in my comment yesterday, that everyone, including me, thought that the problem was with a bad or missing ground, but I had a hunch that you were once again going to prove us wrong. Hahaha. Way to go, Ivan.

[@andrewconroy1266]

Diodes are a very common fix in the electronic access control industry where virtually every door lock or hold back has some form of large coil that generates negative spikes when released

[@leeross7896]

I have been watching for YEARS and I finally GUESSED ONE RIGHT! check the part 1 a diode was my guess.. YEA ME!!!!!

[@mmichaeldonavon]

Ivan, I used to work with DC power supplies that "made" a pulsating DC voltage up to ~100,000 Volts peak DC. We wired 80 diodes, (almost thumb size,) in series, on each of the 4 legs of a full wave bridge rectifier I don't remember the PIV for each Diode, in the string, but it was huge. Those components were submerged in "Oil" for cooling and insulation purposes. Loved working with them. I worked in QC and testing for a company that built precipitators: "NWL Transformer Company," Bordertown NJ." Nothing like listening to the "crackling" of an unloaded ceramic bushing, as you ran the unit to "30% over the AC power input, to the xfmr primary, for 1 minute." This was to see if it would make it or not. :-) Always had my hand over a BIG RED power off mushroom button. :-) Ivan, you demonstrated great brainwork to figure out how to eliminate those negative pulses. Had you not done this, that car would still be "down and out." Great job.

[@nickberetanos7541]

I was thinking you were going to use an external solid state relay to fire that thing but the Flyback diode was cleaner. Well done sir.

[@_RiseAgainst]

That was my thought, but I think you would lose the duty cycle control of the compressor

[@randyr.parker2698]

@@_RiseAgainst Yep!

[@paulravitsky2898]

Ideally the diode should be as close to the source of the spike, ie. the compressor in this case. Excellent troubleshooting and repair as usual.

[@kevin9c1]

Kudos to the folks who were talking about diodes on part 1. I was way off with my front SAM commentary. Although someone did mention that there is a compressor control learn-in feature. I don't think you can "learn" back EMF but would have been interesting to dig into maybe.

[@ruenjou]

Good job. Most auto technicians don't do diagnose like this; they simply remove and install modules because that's also what the manufacturers instructed them to do.

[@nv1493]

I've seen your creative fixes before, but this thinking and solution was exceptional. Wow, well done.

[@rpujol]

Most of the A/C compresors I have seen in vehicles, (even the old ones with manual controls, no electronics) have a built in spike supression diode installed in paralel in the connector wires that feed the clutch coil to prevent those back spikes of the magnetic field from the coil when is disengaged. Good work Ivan, as always.

[@throttlebottle5906]

sometimes they run shielded wire from the compressor and place the diode/s in under hood fuse box, also could be integrated into a relay. been done many ways over the years :)

[@bernardaflores1720]

I have notice that in some wiring diagrams they would not show the location of diodes. My 2004 CRV has some and I found them by mistake. Great solution, Radio Shack brings some old memories.

[@jeffryblackmon4846]

When you showed the negative pikes, I realized a diode was necessary. This is not the first time I have seen diagrams omitting a diode here or there, especially if the load (compressor) is remote from the controllers. Very cool diagnosis, Ivan.

[@NicholasJaroszynski-h3i]

I can verify this works. I used a 2.5 amp diode Thankyou for doing the leg work, It woukd have been a long road to tracing without it.

[@dustcommander100]

A scope also comes in handy on systems with dc to dc converters or switched mode power supplies. Too much ac ripple can cause stuff to go absolutely crazy, and you may not see that on a voltmeter, even if it's set to read ac. Great job, as always, Ivan!

[@davidraezer5937]

Great diagnosis and video Ivan. Also a great lesson to always check with a scope. I would have made the same mistake and probably spent more time and energy until my synapses kicked in. Now I’m chomping on my bit because you didn’t take the cover off of the controller and check for a freewheeling diode in the circuit that might have gone open to ground. Maybe a zener diode would be better suited over a standard diode. I know they used to place them in the coils of some of the compressor clutches. The older Mercedes models that just cycle the clutch have a speed sensor to monitor for belt slip to protect the belt drive. They cycle the clutch off and wait 3 minutes before attempting again.

[@richard3141592638]

All hail the mighty diode from radio shack - what a fix, Ivan. Well done

[@MrOverstuffedcow]

That Radio Shack package brought back memories of my first soldering iron and first second degree burn. Great job Ivan. Your electronics knowledge far exceeds mine. I cook, and I fix computers.

[@yamaha37776]

Epic diagnosis and repair! It’s really a crying shame that part stores don’t have brain cells for sale on the shelf. Awesome video!

[@laurenbush1665]

Pretty much all MBZ of this Era use the driver front Sam module in the operation of ac. I have replaced many sams along with compressors. I am gonna look up that diagram and see

[@PineHollowAutoDiagnostics]

SAM was not involved in this circuit...

[@tracydiller9378]

Great video Ivan as always. Cool Diagnose and fix, can't beat the Scope hands down.

[@geraldsanderson6162]

Yes a comment below said the diode should be as close to the inductive load as possible. I. E. between the compressor case and the control terminal. When the diode is remote, the inductive spike will travel through wires before being snubbed. I would also use a switching diode as they have fast response to catch the flyback pulse. Anytime the words pulse width power control signal is mentioned we should think transients can happen.

[@kellygibson7299]

Scope for the win. I’m amazed at how many applications there are for it.

[@fascistpedant758]

This is fairly common on the Mercedes AC control solenoid with the integrated diode but congrats on finding it independently.

[@MrSparks54]

I'm glad my 1973 450SEL doesn't have all this electronic module nonsense that virtually all new cars have. Awesome diagnosis Ivan!

[@clintprice2123]

Ok, I watched it and wasn’t disappointed as you always nail down the issue! Glad this channel was recommended to me by you tube at least a year ago!

[@guianders]

😂 Ivan you just resolved my charging problem in my jaguar.. never suspected this ground.. thanks

[@citichap]

I performed fixes like this (thinking outside the box) when working on alarm systems back in the day. This brings back fond memories.

[@dweijers2193]

A genius solution Ivan, well done. The original diode (build in) was shorted, the original short !!

[@bradkline4023]

Ivan, you never cease to amaze me. Great solution call, my man.

[@Wayne-p7n]

Yes I remember bernie thompson had a similar problem that was messing with a coil driver in the ecu, thanks to qll the techs

[@ereeves9855]

My thoughts are the wiring harness on the old compressor should be transferred to the new compressor. It might be a bad diode in that harness but it can be tested for that.

[@LesReeves]

There has to be a diode in that plug on the compressor or in the wiring in the dash at another junction in the wiring where ever it is it's burnt out the poor old DIYer would never have found it & most won't spend the money on a scope. They (like me) would have to find someone like you to effect a repair out here in AUS the problem is finding a trusworthy tech .Thank you for this it has been an eye opener.Cheers mate be seeing you in the next video.

[@brianw8963]

There sure are many very intelligent techs on Your channel and My hat is off to all of them when it comes to this more intricate electronic field, but it is amazing to Me how You tackle and sort this stuff out in short order. I know in My local neck of the woods, this owner would be up the creek without AC. Nice work. 👍👍🇺🇸

[@RK-kn1ud]

99% of mechanics aren't Electrical engineers. Most mechanics are far BEYOND average intelligence when it comes to mechanical issues, but they are only slightly above average intelligence when it comes to semiconductors. I highly suspect that Ivan understands more about silicon than your neighbor... probably because he has had much more education in that field. He also seems to bleed oil, which makes him a double threat.

[@dans_Learning_Curve]

​@@RK-kn1udLoL 😂😆 When he got tired of the 9-5 and finishing grad school! Yes very talented in many ways!

[@johnjennings2672]

Wow! A diode! I just serviced a fuel injection ATV for a worn out clutch and poor running motor. Replaced the clutch first then checked the motor. Would start and idle but no power and stall under load. Nothing seemed wrong, but a Google search said check the plug type "ATC" diodes in the fuse block. Found 2 of the 3 reversed. Switched back and engine ran like new!

[@akdenyer]

Hi Ivan I use a lot of relays in my designs and they should always have a back emf diode when being fired from semiconductor drivers. Some relays have an internal resistor to reduce back EMF. Relays and solenoids are effectively the same thing. Coil of wire. The diode should be fitted as close to the inductor as possible. Helps reduce RFI. But as you say where is the diode in the replacement unit? On AC circuits I use an MOV.

[@Graham_Wideman]

Mostly agree for DC. For AC MOV will address infrequent high-voltage transients, but in this application a transient will occur on most turn-offs of the inductive load. MOVs degrade with each transient, so their lifetime has to be factored in. An often user alternative is the "snubber" circuit" using resistor and capacitor.

[@paulsolovyovsky1702]

$.25 part fixes a thousand dollar issue. It's not about the parts but understanding the problem and getting to the root cause. Great job

[@Sandmansa]

That was a nice little hack to get the results you needed Ivan. Well done. Your customer will be very happy that the A/C is working once again. But there has to be something else we are overlooking. Because not everyone will have easy access to electronic components like that.

[@rj.parker]

Anybody can buy a diode. Expertise is hard to buy.

[@Sandmansa]

@@rj.parker True. You can still buy electronic components. But it's basically only mail order now. And I agree with how hard it can be finding a great tech with a good wealth of experience.

[@PineHollowAutoDiagnostics]

Is there a global shortage of diodes? 🤔

[@tschuuuls486]

The thing that "we are are overlooking" is that the replacement compressor is not build correctly from the factory (missing internal diode). Replacing the internal one there would mean: 1)evacuate the refrigerant 2) remove compressor from car 3)replace compressor for warranty (they might not be able to reproduce the issue at this step) 4) reinstall 5) recharge AC. Soldering in a diode is way easier and cheaper.

[@Sandmansa]

@@PineHollowAutoDiagnostics Not that I am aware of. But there is a shortage of local brick and mortar stores that sell them.

[@randyr.parker2698]

The standard diode used in electronic applications like this on relays is a 1N4148, has thinner leads, easier to twist around that wire, but I understand, you use what you have available at that time. Good Job!

[@Graham_Wideman]

1N4148 is a signal diode and not suited to this task. Perhaps you means one of the diodes in the 1N4001 through 1N4007 family?

[@JW-ce4lx]

Best place for the diode would be at the compressor. when the magnetic field in the coil that operates the valve collapses, that is whats causing the negative spike.

[@russellhltn1396]

Yes, typically diodes are installed at the source. As for a retrofit, I would have looked at putting it inside the controller. It would have been better protected against future techs moving things around.

[@calopii]

Yes, this thing is typically called flyback diode and normally mounted close to the inductive load. See here: en.wikipedia.org/wiki/Flyback_diode

[@xalty1200]

the diode is on the vd valve aftermarket pos compressor didn’t have it

[@Stoney3K]

@@russellhltn1396There's no harm in having the diode installed in both places though. The closest diode will take the largest amount of the back EMF current.

[@blackcat31w]

I too think the diode should be as close to the offending coil as is practical. By installing the diode at the controller we are allowing a current spike to travel along the wire in the harness then on the "fat" brown wire and finally to the chassis and back to the compressor. The controller does not see any kind of spike but some high impedance circuit wire (like a hall effect sensor) in the harness might pick up some EMF from the current spike.

[@fordtruckguy5545]

A bad diode was the problem on my Ford pickup with the clutch not engaging. My co-worker and I beat our heads against a wall, trying to figure it out. In the short term, we had run a removable ground wire to the body that I would reconnect for the short drive home each day, for a week or so. I found the diode fix on a website forum and it fixed the issue. The funny part was that the factory OE diode was AC Delco, my local Ford dealer didn't carry it in stock, but the Chevy dealer had it.

[@chrisgiddings7314]

Amazing diagnosis and I bet super satisfying when it worked amazing fix

[@roryeilers2451]

outstanding video series,, this is actually a hugely notorious and well-known issue,, for the clutchless compressor mainly found in luxury vehicles like this, there are two control valves available depending on application: one with an integrated diode and one without-- yellow dot on the plug indicates diode present. however, both genuine and aftermarket compressors alike are arriving DOA from the factory, fitted with valves either missing the diodes entirely OR, diodes of inferior quality and/or incorrect rating OR, diodes wired backwards or faulty solder joint or bent pins etc. you'll deliberately order a compressor with a valve with the yellow dot, or just a valve alone with a yellow dot, and it still may not work because quality control has slumped so hard in the last decade. scant few have ever even heard of a clutchless compressor-- horror stories abound of owners spending THOUSANDS on all new ac componentry to no avail, OVER A $.10 DIODE. the shops and diyers that are actually successfully servicing these things have taken to soldering a redundant diode into the control wiring exactly as youve done, most often right by the valve, or up top in the harness, grounding right to the valve cover. again, wonderful videos, really shows in depth whats actually going on electrically and electronically here!

[@OzFrog48Z]

Maybe just for comparison you could just plug in the old compressor and hook up a scope to compare the signal with the new one. Possibly the old one had a mechanical problem, not an electrical one. I've scoped the pwm signal on a few compressors and have never seen a negative voltage spike. They must have a diode somewhere. Good find.

[@carltonlane8931]

Lesson of the day,we need to use scanners,understand the information,somewhere in that Mercedes there would have been a diode,probably on the original compressor or bcm,without the scanner ,you would not know about the return voltage,bloody good video,thanks for the video.

[@PineHollowAutoDiagnostics]

You mean OSCILLOSCOPE? Very different from the SCANNER 😉

[@RK-kn1ud]

I am surprised there is no flyback diode displayed in the OEM diagram. This is the kind of issue that the dealer would fire the parts cannon at for 10-20 visits and then want to sell you a harness replacement.

[@hightttech]

Diode probably integrated with controller or OE solenoid, thus it doesn't appear in diagrams as discrete component.

[@xalty1200]

the dealer would’ve fixed it on the first visit because they don’t entertain junk parts

[@VidelicetMoi]

I do not like this "tacky" third world fix. I would have demanded as an owner the proper exploration of why. Why didn't replacement parts not fix the root cause and how do I return the state of operations to OEM. This third world mentality is sub-par. This is the fault of the owner-- he wants an expensive car strictly for show but not the expense for doing it right.

[@RK-kn1ud]

@@VidelicetMoi If you can explain why a flyback diode is required for an inductive circuit in the context of this repair in a manner where you are teaching me how this is "tacky 3rd world fix" I will entertain your opinion. Maybe you might teach me something, and I'm always willing to learn something I don't know.

[@RK-kn1ud]

@@hightttech Buy why would the replacement controller exhibit the same issue?

[@htownblue11]

Slick fix there Ivan. My guess is there was / is a diode that is broken because you tried two compressors and two climate modules I believe. If a non functioning diode was present in either and replaced my guess is it would have worked. Bet they had a diode in the harness or some such thing. Regardless, excellent case study.

[@rj.parker]

Typically the suppression diodes are in the controller. Probably was in the compressor in this case with the original failed and the rebuilt unit missing it.

[@sprograt]

Ive been really looking forward to part 2 of the Mercedes GL air conditioning fault, Thanks Ivan.

[@Santiago-je5ki]

WOW, just amazing. Ivan whats your origin story? Did you start in the component industry, how did you get into the automotive industry. Brilliant work.

[@tomtke7351]

In my design arena we did lots of relay switching and the switch was transistors. Conveniently, the relays we used had built-in diodes ACROSS the relay's coil. These were D.C. voltage coils and the diodes were such that they were reverse biased when control voltage was applied. How does this work? In operate mode the current flows from battery to coil to ground back to battery. In open circuit the current then flows thru coil in same direction thru diode back to coil. A coil builds up magnetic flux when the current flows. When the current's source is switched off the magnetic flux collapses. This causes the voltage across a coil TO SWITCH POLARITIES. Voltage applied to a coil causes the coil to consume energy with a resulting voltage drop across the coil (+) to (-) from input to output. When switched off, the coil becomes a generator by switching voltage polarities to continue current flowing; now (-) is (+) and (+) is (-). The flowing current wants a path thru the now open switch. An inductor 'resists CHANGE in current flow.' If a switch opening stops current flowing, the coil then goes crazy [THE CURRENT WILL FLOW somewhere, somehow] and so it first flips polarity then INCREASES IN VOLTAGE looking for a path for the collapsing coil's current. This is a cause of pitting of switches because the extremely high REVERSE voltages cause arcing across the switch's now open contacts. It's my belief that early vacuum cleaners had headlamps across the motor to provide a motor coil discharge path when the switch was turned off = opened to eliminate switch pitting. In Ivan's compressor the control voltage was applied to the compressor's coil, thru the coil, then to ground. His diode fix then provided a diode path from ground to the control wire which goes back to the compressor's coil. It is a longer route but sufficient. I'd of preferred a diode across the compressor coil for a shorter path. Internal to compressor is best. Another responder in part 1 suggested certain compressors had such diodes; that some did not. I'm uncomfortable with a 1amp diode in a circuit commonly handling 1amp frequently... I'd want at least 2amps. There's derating criteria to include temperature, voltage, end-of-life, etc., to precisely derate a diode's current specification for its application. COMPRESSORS get really hot.... being inside the car is a less severe environment. Still 1amp for 1amp application is cutting it close. Keep in mind that in addition to the other currents flowing in the ground wire Ivan used there's now an additional 1amp added. Is the wire guage sufficient? For the electric techs: ohms's law for inductors: V = L×(di/dt) Voltage developed across an inductor (neglecting its resistance) is related to the CHANGE of current/over time) Increasing current the inductor's voltage is (+) -> (-). Decreasing current inductor voltage is (-) -> (+)

[@PineHollowAutoDiagnostics]

Diode package said 3 Amp lol

[@tomtke7351]

@@PineHollowAutoDiagnostics 3amps is V E R Y GOOD! Still ground wire amperage is important too. And there's numerous switches in my old 90 Buick that should have coil diodes but don't and contact pitting is evident. One door switch actuating four door lock coils is significant. All of us need to use o'scopes more often. A weath of opportunity to more quickly get valuable info.

[@peterlast3200]

I thought that was very clever regarding the "final solution" whilst working on a Mercedes.

[@Pdubforever1]

i literally have this exact same problem on this GL I'm dealing with. how do i oick the correct diode? i was almost just about to pull the trigger on a hvac module. im glad i saw this video.

[@jgeorges3061]

Ivan, AWESOME WORK as always thanks for sharing.

[@mdaa2kv2]

Looks like the diode is in the old compressor's connector. This is why you have two wires from the body of the compressor going to the connector but only one pin to the car harness

[@r.weaver3769]

Mercedes uses press fit pins and super brittle lead free solder on connector pins. It's why most modules fail in MB cars. Re-working all the pin connections with regular solder on soldered and press fit pins fixes many failed Mercedes modules or prevents future failures.

[@dri50]

I have a new Ford Bronco Sport. These still have a magnetic clutch on the compressor. The schematic shows a diode built-in to the magnetic clutch. So a diode is almost always required to suppress the negative spike from a coil, whether it's a relay or magnetic clutch.

[@clintprice2123]

Nice I was just looking for something worthy to watch and you posted this, great timing Ivan!

[@stevelang3171]

nice! a freewheeling diode to clamp the back EMF from the compressor coil. radio shack used to make shrink tape to use when one forgets/ has no room to put shrink sleeve on.

[@ThunderbirdRocket]

That was an AWESOME lesson ! Super interesting problem / series of challenges you confronted and handled here ! It’s a great day when another one of your videos comes out !!

[@robertheim352]

Great advice about using scopes. When properly connected the scope shows the detail of voltage over time. But, you surely realize that interpreting the signal is the other facet of analysis. (In this case: modulated control, sequenced power up and reverse EMF.) I like the way you converged upon the cause, experimented and resolved it. Question: Did the ebay module work as well with the quenching diode installed?

[@aranhaydar9195]

Ivan, This was by far the coolest video of the year 😎

[@MR.KNOWITALL1982]

Ford uses flyback diodes built into the A/C compressor connector. Not sure about other manufacturers. You can actually puchase them diode connectors after market. Nice work!

[@davyarthurs]

One of your most impressive thinking outside of the box fixes so far - for me anyways. Skilllzzzzz

[@PineHollowAutoDiagnostics]

🤓👍

[@Stoney3K]

Super work Ivan. Keep in mind, the best practice is to place that flyback diode as close to the solenoid as possible, because the wiring acts as a transmission line and the high-frequency current of the spikes will flow between the coil and the diode -- you want those wires to be as short as possible and shielded/isolated from any neighbouring control wires. I am really surprised that the flyback diode isn't integrated with the valve solenoid.

[@Graham_Wideman]

Well, not a transmission line, but rather an antenna.

[@Stoney3K]

@@Graham_WidemanWith regard to other wires in the same loom, it will act like a transmission line, as those wires are capacitively coupled. Which will cause high frequency crosstalk.

[@Graham_Wideman]

@@Stoney3K Ah, I see what you's saying. Yes, definitely there would be inductively- and capacitively-couple crosstalk to other wires in the loom. But in my field we might not really think about them as a "transmission line" per se (in the high-frequency sense, not power-line sense) unless the situation fulfills some other requirements that make it amenable to "transmission line theory". (Two specific conductors carrying a signal from point A to point B, conductors consistently spaced so as to allow summarizing their incremental R, L and C as bulk "impedance" value, no significant random influences from other conductors, and so on.)

[@shuriKen469]

another banger of a diag and repair!! the Final Solution for the AC will make for carefree roadtrips in German MB.

[@reubenmiller7988]

It takes real knowledge and experience to pull off such a cool repair

[@richb419]

Normally the diode is connected across the inductor as close to the inductor as possible in a normally reverse biased condition, that shorts the spike that is generated, like when you see relays connected to a micro controller (freewheeling diode). the diode must have a very large piv compared to the working voltage because the negative spikes can be substantial. good call Ivan Rich

[@kevinshirley4598]

Well done!! great video. Could you have place the diode in series to ground on the ground leg?

[@ChrisVargasrpht2000chris]

That is great that you fixed it but what really happened, did the customer buy an aftermarket compressor with a missing diode, did the original diode go bad. What are your conclusions on why it stopped working.

[@rj.parker]

He gave his theory. Watch again.

[@gregwarren8583]

That is a spike wupression diode that has been used since at oeast the early 70's (GM put it across the clutch connector - soldered in) and taped up on the end of the connector. Repaired many over the years. Diode test on the DMM, or old analog meter would show the fault.

[@rj.parker]

The lesson was to use a scope especially on variable controls

[@faxmen09]

Lord won't u buy me a Mecedes Benz...Not. I want to know where the diode that went bad was/is located? When comes to electrical issues, imo Ivan your electrical expertise, second to none. Nice work, again.

[@confusedbuck]

Very impressive fix. We'll done

[@lawrencecavens5760]

Ivan!! that is sew kool/ Yeah I know bad dry joke. BUT to me it sounds like there was a manufacturing flaw from the factory where either the diode was too weak and or OR the control circuitry was not manufactured properly and a step was missed - In any case YOU Fixed! IT... Wayda go there Cool DUDE!!! The one thing that can only cause this is unregulated or poorly regulated charging of some sort - Boosting battery from jump starting or bad battery charger with a bad A/C ripple or Welding without disconnecting battery first. WOW talk about Delicate electronics ...

[@westtexas501]

"Hey hey hey" ! the sound of victory lol. Good ol RadioShack to the rescue. unfortunate they went out of business.

[@mikedaugharty5544]

Better schematics would have been good to show that diode in the control Circuit of the compressor but you figured it out anyway and the rebuilders should pay more attention to what they're doing and do it right but until things have more more quality controls behind their processes and the schematics of the corporations we will have to fight all this stupid stuff but you did a bang up job there awesome awesome awesome Ivan

[@190055joe]

Good job but normally to stop back EMF you install the diode in parallel in series the diode could get hot and you have 0.7V drop.

[@gerrymoeller]

In most cars I've worked on this was actually how the manufacturer hooked them up. On older GM cars It's usually across the compressor plug it's self, others I've seen in the fuse box..

[@rj.parker]

He wired it in parallel.

[@gerrymoeller]

yes that's how GM wired them. Across one terminal to the other and used the crimps to fasten them.

[@hightttech]

He did install in parallel; the anode is going to ground. [thanks to @Graham_Wideman for bringing to my attention an error in my comment 🤦; i fixed it]

[@breezy1849]

He did install it in parallel. The control wire goes to ground inside the compressor and Ivan connected the diode between control and ground at the controller. Draw out the circuit to check the installation.

[@petepeabody8905]

Wow great, but can you verify that there was in fact either a bad diode in factory harness, or the aftermarket compressor being the problem with no diode?????. That would be nice to know as well.

[@mykofreder1682]

Additional test I would have used the compressors ground in the test, went over 1 amp in bulb testing, and voltage scope of the bulb ground, compressor ground, and compressor no bulbs.

[@michaeldamian5140]

Very clever; impressive diagnosis and repair.

[@tomvleeuwen]

In this case, I have to agree that I can't think of another method to determine if this diode is present than using an oscilloscope. The fact that the diode should be somewhere in the circuit can be assumed, since it is PWM'ing an inductive load, but since the schematics don't show the diode, there is no way of knowing if the diode should be in the compressor or in the controller. If you would know that it should be in the compressor, you could measure it by applying a 500mA reverse current and see what the voltage is. The diode would clamp it to 0.6v (You could still do this with the old compressor) But if the diode should be in the controller there is no other method than using the oscilloscope to see if it working.

[@alexanderzubar9593]

Reversing a limited current through the coil is the only way you could tell if there is a diode in it. An easier way would be to apply at least 2 volts with a 1 amp limited supply. This would be enough to show a diode (0.7 volts) and any fly back diode on a coil like that would handle 1 amp.

[@cullenmiller8170]

Nice video and diagnosis. Here I am trying to sell my Snapon Verus with the 4 channel lab scope. 🤔

[@neilmurphy845]

Boom it's finally fixed you're genius Iven

[@fluffyblue4006]

The best place for a fly-back suppression diode is close to the coil that causes the back-EMF spikes, because the conductors between coil and fly-back diode serve as antennas that radiate an RF signal. This is PWM, so you'll get a signal with a base frequency that might be in audible range, with a lot of harmonics on it that might interfere with radio reception. With this repair with the diode in that location, one should test the radio as well.

[@Graham_Wideman]

Yes, I agree with this assessment. It would be good to try an AM radio near the wiring to check how that varies at different AC settings. Wait -- who has an AM radio any more?

[@fluffyblue4006]

@@Graham_Wideman My car radio has AM, FM with RDS, digital radio, bluetooth, car-kit, aux in, audio file playback via usb. AM still offers a lot of news and talk radio.

[@Graham_Wideman]

@@fluffyblue4006 Yes, I was kidding about the AM radio. I myself have AM in both my vehicles. Although AM is starting to be dropped on EVs, I have heard.

[@fluffyblue4006]

@@Graham_Wideman They're dropping it because the EMF of those power electronic drives will be too much and too hard to shield properly.