Very Cheap and Easy Mod to Fix the Horrible Ripple, Below 1mV RMS ! - Lavolta BPS305 MKII

[@benfranklin4524]

"...shell shocked by the magnificent idiocy he has just witnessed..." This phrase alone made this whole vid worth watching.

[@GoatZilla]

Management: "It's cheaper to just fire the engineers. Bonuses for management this year!"

[@TickertyTonk]

Thank you, thank you, thank you. My Lavolta output ripple went from 112mVp-p to less than 3 mVp-p just by fixing the grounding.

[@ericrawson2909]

I just built a linear regulator and was disappointed by a few millivolts of ripple. Eventually I realised that the earthing wire with a croc clip on my scope probe was lying next to the transformer core. I moved it away and the ripple disappeared. The lesson is that you should keep sensitive circuits well clear of transformers. Stray magnetic fields cause induced emf. The transformer even has a flux band to reduce stray fields.

[@233kosta]

I have a similar issue with a cheap Sony turntable. The ripple is audible. At first I thought the filter cap had gone, but no. Replaced filter cap - ripple's still there. Eventually I realised that the transformer is just way too close to the amp circuit board (which... why does a turntable you sell with component systems even have that, but whatever). One day I'll finally have nothing better to do and move the transformer.

[@Patrik6920]

@@233kosta u have any model number for it? (i hope it may be possible find a pictures of the inside) ... it may be simpler just make a simple shield and/or filter caps .. most likely thay didnt add any filter on the primary -> ground noise is being amplified.... could test wrapping the power cord around a ferite atleast 2 laps and see if it cancels out.. if it does u have could have a ground noice problem.. (cused by something else conncted to the electric system/ground...motors are one huge source for such noice... washing machine etc..

[@carlosanvito]

Great fix. The source of noise was likely due to eddy currents flowing in the transformer laminations causing a ground loop via the ground wiring. Rule number zero: ground everything to a single point. Treat that point like Rome (that is, all roads lead to Rome). Any other arrangement is certain to introduce noise - you elegantly demonstrated that.

[@lmamakos]

Good on you for scraping the paint away from the bottom of the case!

[@mk6595]

This was useful. I opened up my Korad power supply and found a similar situation, that I'll soon fix.

[@paulcohen1555]

You are VERY LUCKY to have the old, linear, not switching type of power supply.

[@Citronitroify]

"Magnificent idiocy" - You got that right!

[@siberx4]

What a great fix and fairly easy to do. As you noted in the last bit you also scraped the paint below the grounding foot of the transformer to ensure good contact (a great move that they _also_ should have implemented in the first place); I wonder if even the original wire routing would have performed nearly as well with proper electrical contact to the casing at _both_ grounding points; this would have shorted the current generated by the transformer plates, and the steel casing should have provided a much better ground path than the transformer body and would have cost them even less than wiring it properly (although they should have done that too).

[@helifynoe9930]

Way back long ago, I purchased an expensive Sony TV, but the picture quality did not seem that impressive. If you were to say have a complete red picture across the entire screen, it instead was red for about an inch on the left hand side, and then the rest of the screen was pink. So I removed the back of the TV and slid out the main PCB since it had this slide option. Then I turned on the TV and now it had a full proper red screen. Push the board back in, and the trouble was back. So I monitored the 12 Vdc in both positions and noticed that when the board was back in, there was an extensive ripple on the 12 Vdc line. I then realized that the CRT yoke was the source of the problem. Once located closer to the PCB, components on the PCB were picking up the horizontal signal. So I connected a wire to GND, wrapped it once around the yoke, and then connected the other end to the 12 Vdc line via an electrolytic capacitor. Boom, problem gone. This now cancelled out the pickup. I informed the store where I had purchased it from, and they pulled this model from the shelves and sent them back.

[@robc4191]

Great content. After watching this i went and purchased one of these power supplies, but when i received it, it was a switch mode model instead! Sometimes you just can't win.

[@MatSmithLondon]

Shame - and the data sheets are rubbish so it’s hard to tell!

[@MatSmithLondon]

Dear K Tech. We at Lavolta take our customer experience very seriously and will be reviewing the design decisions made here. In other news, how much did that bit of earth wire cost, can you still feed your family?!

[@wrrr10g53]

Nice instructive video, thanks for sharing. Even greater storytelling after 4:49, LOL😄

[@k-tech796]

Thanks for watching, I'm glad you enjoyed it!

[@mikeb6185]

These sort of discussions/design decisions happens all the time in IT too if management and PM’s get involved 😂.

[@DonzLockz]

Haha, hilarious and great job repairing it. The really need to make sure the Asian kids are over 10 years old who work in the company and manufactiring line.😂👍

[@erikhovdahl]

I had the same problem , nearly 45 years ago. Did build my oven transformer to a powersupplay when i was sin engineering class. Solution was to weld a seam over all the metal plates of the transformer. Connecting them solid together. Reduced the 50 Hz noise.

[@BlondieSL]

Now to play with the switching power supplies that look just like these. No huge Xfrmer, so not heavy. But they can be a real pain sometimes. Especially when powering up. The spike can be huge and blow up equipment if not careful.

[@MrLoyssa]

I also expected this problem at first, rather than a "ripple" of a few mV.

[@friedmule5403]

Great video, this make me wonder why the very expensive top-brands often have several mV in ripple? :-)

[@Robert-hr6sh]

One thing I would have done as well. Good video to show how manufacturing techniques are also short comings of what needs to be done in thinking as what should have been done, bottom line. 👍 remember to be well grounded in the design first.

[@joseluisvaiksnoras7857]

Single point grounding is a must.

[@Ratkill]

"Shellshocked by the magnificent idiocy" "and they are already whackin eachother off" Now this is real literature

[@smccomas01]

You have listened in on management and investor meetings.. subscribed :)

[@catalinalb1722]

Close to reality... thanks to the video and explanation! By the way if the GND is connected to the Negative (COM) will in general add or cancel the noise ? We always connect those two together with a short cable right at the PowerSupply (Mean Well) terminals. I notice noise and want to get rid of it. Will try by removing that bridge.

[@rolandburisch9489]

Good video, thanks. But it occurs to me that you may have improved the ripple further by wiring the output ground terminal directly to the ground input terminal on the mains IEC connector. If there's any leakage in the transformer, then a very slight volt drop would appear across the ground wire at the grounding point on the transformer mounting screw. By connecting directly to the power input socket this could be eliminated. Either way, your final ripple voltage is good and very much better than it was.

[@gearscodeandfire]

I thoroughly enjoyed this video that the algorithm decided to pop up for me, I learned a lot

[@k-tech796]

Thanks for watching, I'm glad you liked it. I'm also baffled why the algorithm decided to do that now, three years later...

[@axelr.8766]

Isnt your oscilloscope also grounded? Doesnt that introduce further path of current flow, which needs be taken into consideration?

[@NackDSP]

Ground star configuration for the win.

[@etmax1]

From memory the safety earth/ground wire has to be rated such that 25A can be sustained for 1 minute without the wire fusing or getting unsafely hot.

[@bobreiber2066]

The design was OK but the implementation left a lot to be desired. I wonder if the even did any ripple measurements on the assembled product.

[@matthewsheeran]

So always connect your grounds directly together! I might also suggest mid-frame perhaps. A dedicated ground mount point too and definitely not off a transformer for example. Oh the cost the cost geez! Clearly the cheap Charlies are all in China.

[@233kosta]

At this point I consider myself fortunate to be working for a business that doesn't have senior management Muntzing products for us.

[@k-tech796]

You are fortunate, working the management is my least favorite activity.

[@233kosta]

@@k-tech796 The business went private just before I started. I'm told it had been pretty nasty in the lead-up, while still owned by hedge funds/etc.

[@reylambarte5615]

I think the main cause of the ground problem is the lamination of the transformer core and the paint on the chases

[@lohikarhu734]

normally, i would expect a lab power supply to be 'floating', with 1000 V isolation from the 'earth' ...if it's not 'floating', you can't do things like connecting in series for +/- supplies, and even in parallel, due to 'earth' differences.

[@windward2818]

Your assumption that there is a competent engineer looking at the overall design performance is very optimistic. In the USA with electronic instrumentation design, which would include a bench power supply, the modification would have the rear and front panel earth wires go to a single crimp eyelet, which would then be attached to the transformer grounding hardware (case hardware). The eyelet barrel size is chosen to accommodate both wires with a single crimp using an approved crimping tool. This way if the case connection (transformer case hardware) comes loose you will not lose your earth ground between the rear plug and the front connection. To put it another way all conveyed earth connections are always mechanically rigid without depending on screw hardware or a soldered connection. If a soldered connection gets hot it can come loose. So, the rear ground connection would be a crimped female crimp terminal (usually insulated) designed to fit the tab on the power inlet. Which would also be crimped with an approved crimper. For a complex ripple noise waveform many times it will be specified in RMS, which implies the measuring device be true RMS with a rather large crest factor and a large enough bandwidth. Many newer oscilloscopes do a very good job at mathematically calculating the RMS value of a complex waveform (with a large enough sample size). But, it can be worthwhile to get an FFT of the noise, and also in some cases THD.

[@yeyuan6273]

woooooo, its not isolated, be careful!

[@aurthorthing7403]

Cool observation and thanks for sharing. I just bought a Dr. Meter branded one of these and I bet it has the same issue. I plan on taking it to bits when I get over a migraine.

[@k-tech796]

Thank you for watching, hope your migraine is better now.

[@aurthorthing7403]

@@k-tech796 thank you. The Dr. Meter branded PSU is built the same way. The hope of lower ripple was the reason I bought a linear power supply. Thanks again for addressing this issue. Since these are very popular bench PSU I would imagine more people would also like to know about this issue and modification. And yes, migraine is much better now.

[@scriptguru8]

Very useful finding, but couldn't the same mod be done in a few minutes without fast forward?

[@phuang3]

780uV RMS for this kind of cheap power supply? That's insanely good.

[@RingZero]

Thats not a surprise, there is no switching - this is a linear PS

[@phuang3]

@@RingZero Well, my old benchtop linear power supply come with 2mV rms noise and 20mVpp ripple.

[@vitacell1]

Thanks, useful info!!

[@k-tech796]

Glad it was helpful!

[@foobarables]

It's not clearly visible but the mains wiring going underneath the printed circuit board should be double insulated.

[@ntal5859]

Clearly they were never taught about star configuration for grounding.

[@olafschermann1592]

Great work

[@k-tech796]

Thanks, I'm glad you liked it.

[@imqqmi]

Why would you want to connect earth ground to the negative output terminal?

[@iwantagoodnameplease]

Do you have links to your channels on library and bit chute?

[@k-tech796]

Small creators like me are not allowed to post links on KZbin. But the channel and video names are the same, so my content is easy to find there with the search functions.

[@colloidalsilverwater15ppm88]

Great work. Almost like detective.

[@k-tech796]

Thanks, much appreciated.

[@str0y]

What if I don't have a face ground connection in my electrical installation? Does the video still work?

[@k-tech796]

Probably, I can't think of a reason why it shouldn't.

[@hifiboy3896]

Klasse Videos! Danke. Ich werde mir dieses Modell holen, da es für meine Zwecke ausreichend ist. Ich mache viel mit Vintage-HiFi und Kofferradios als Restauration für meine Sammlung (+ YT Channel) und da kannst du kein "neuartiges" Schaltnetzteil brauchen. Dein ganzer Lang- Mittel- und Kurzwellenempfang wird sonst verdorben durch diese massiven Störimpulse. So kannst du nichts reparieren oder abgleichen. Dachte so etwas klassisches, lineares mit Längsregler gäbe es schon nicht mehr neu zu kaufen heute. Gut zu wissen :-) dazu sind die 3 von dir genannten "Mods" rasch und günstig umsetzbar. Habe mir auch deinen Vergleich zum "MK II" angeschaut. Problem ist es dann das MKII eindeutig zu identifizieren. Selbst wenn du genau hinschaust weißt du nie ob du dann final v1 oder v2 geliefert bekommst :-( Auch der Belastungstest (sehr schön dargestellt) über die Zeit mit Erwärmungsmessung war sehr akzeptabel insgesamt vom Ergebnis. Ich brauche zwar oft nur wenige 100 mA, aber es ist gut zu wissen, dass bei Bedarf auch mehr geht. # English # Great videos! Thanks. I will get this model, because it is sufficient for my purposes. I do a lot with vintage hifi and portable radios as restoration for my collection (+ YT Channel) and you don't can use a switching power supply. Your whole LW, MW and SW reception will otherwise be disturbed by these massive interferences. So you cannot repair or adjust anything. Thought something like this classic, linear with longitudinal regulator would not be available for purchase anymore. Good to know :-) in addition the 3 "Mods" mentioned by you are fast and inexpensive convertible. I also took a look at your comparison to the "MK II". The problem is to identify them clearly. Even if you look closely you never know if you will get final v1 or v2 :-( Also the stress test (very nicely presented) over the time with heating measurement was very acceptable overall from the result. I often only need a few 100 mA, but it is good to know that more is possible if necessary.

[@SusanAmberBruce]

Very funny musings Ha ha ha and great video thanks

[@k-tech796]

Thanks, I'm glad you liked it.

[@jonnymakers9560]

I never use AC coupling because I want to see the entire vdroop. I always use DC coupling so I can see everything. I see a lot of folks using AC coupling but..the data is limited

[@Andrew-dp5kf]

I think it’s more likely that they had a warehouse full and discovered the issue, the solution as they saw it was to just remove the ground bar. Technically rubbish, but very pragmatic you have to admit.

[@k-tech796]

Yea, it's sometimes simply astonishing which creative routes people can take to get it done.

[@JurekOK]

To be fair, if they made it with the correct wire, it would not show on the top of your list when you "sort by price". Managers are there for a good reason.

[@HitAndMissLab]

Great video, thanks.

[@k-tech796]

Thanks, I'm glad you liked it.

[@alesizzz1]

Well done

[@norbert.kiszka]

Cutting corners? Then why You are using that cheap soldering iron?

[@Peter_Riis_DK]

If you draw one amp through a bridge rectifier, all things equal, it dissipates around 2.8 W. If that energy has nowhere to go but the ambient air the high temperature isn't surprising. Btw. it seems you haven't uploaded in over three years - why?

[@k-tech796]

Even small things can make big problems. Making these videos is a lot of work, at least 2 hours per minute video runtime in preparing the topic, shooting the material, editing and posting. And unless you go viral, you get only a couple hundred views. But I think I was close to finishing the next video, so I should be uploading a new video soon.

[@Atetus94]

Very interesting!

[@richardballinger517]

Better to take all earth wires back to the earth socket or single fixing point, don't daisy chain them.

[@matthewhopson964]

Many thanks.

[@bens4446]

Hmmm I see the same ripple waveform on my DSO-138.

[@ec11368]

Thanks a lot

[@Cinncinnatus]

simple add a capacitor to your output. problem solved. Go build a AC/DC full bridge rectifier regulator and you'd know its simple as that. 3 Farads does wonders for elimating ripple. (no I didnt typo that... yes 3 farads)

[@k-tech796]

Preferably build out of 10,000 low ESR caps for that ultra peaky kilo Amps short circuit current, for when you really want to blow stuff up ;-)

[@hqiu6828]

The transformer should be properly shielded to prevent interference with other devices nearby.

[@garydirkse9900]

You should not have to fix new equipment. I like the large transformer. A very good fix but, too much work.

[@lesliepieterse1072]

There is almost nothing more to add to this comment section as the narrator has done so already 😂

[@pibbles-a-plenty1105]

It is contrary to safety standards for a protective ground to be connected to a point that's primary purpose is other than safety related. The transformer mounting bolt must be unfastened to service the transformer. A safety ground connection is then compromised, it's no longer "safe." The mains power cord green/yellow protective earth wire must be connected to the chassis with it's own grounding point with nothing else using that grounding point. So, Mr. Smart-Ass K-Tech, you are WRONG! The GROUND POINT should NOT be at the transformer mounting bolt. Read the safety standards, ISO, UL, CSA, BSI, etc. ..............

[@k-tech796]

Good point, I'll keep that in mind.

[@tamask001]

It isn't a great idea to solder the ground wire, because if current starts flowing on it, it may desolder itself and let the human conduct the current instead. It should be crimped instead...

[@k-tech796]

Good point. Although it might not be always that easy to implement, but certainly well worth a try.

[@lohikarhu734]

this measurement is a bad problem, as it makes no sense to be using a chassis ground for your "negative"...without a truly "floated" scope, you have a multiple ground problem to deal with, and this is just bad practice...unless you connect the ground on the scope to a "hard" ground on the power supply, there's often a significantt "ground"voltage differential...in some "industrial" cases, you can have several volts difference, with a LOT of current possible..

[@ABaumstumpf]

"How to fix your ground-ripple: This 1 easy trick that PSU-manufacturers do not want you to know!" a 10cents fix is kinda disgusting to see. If the bad ripple just reduced their sales by 0.01% than had cost them more money than the wire.

[@aleksandartomic5515]

Did not have time and patience for that, just binned

[@dell177]

Most managers are defrocked accountants and should be kept far away from engineering decisions.

[@oldskeptic1513]

... here we have a peace of cheep Chinese electronics, where R&D are not part of the business model and expense... just like quality control is neglected ... it's not that they do not know, it is they do not care ...

[@k-tech796]

Yea, it's the "cha ba duo" mentality all the way.

[@engjds]

Chinglish think star earthing is an American sci fi drama. You take a look inside an old Hameg class power supply from the 1990's, there are a billon more components, with electronics, morer is betterer.

[@alevans51]

Linear is better. :)

[@alfabethev2.074]

Well . That meeting was of course held in some 💩chinese language...🤷🏾

[@ahndeux]

I think you're just overthinking it. The reality is the mechanical engineer has no idea what ground means. As long as its connected to a conductive point on the chassis, its ground. Case closed. I'm sure the electrical engineer is also some new guy who doesn't know any better. They probably never heard of a ripple.

[@josephkaminski1857]

LOL what a fool. All he had to do was clean the paint off the chassis where the transformer legs bolt so that all the transformer legs made good equally conductive contact to the chassis. But all he did was destroy the isolation of the chassis from the NEG terminal by making them no longer isolated. What a fool. You should never really connect the GRN to the Neg terminal on the front anyway. The ground terminal is there on front so you can connect it to whatever unit that is under test chassis, so that both the power supply and the unit under test have the same chassis ground. . Chassis GRN is where all the filtering caps dissipate the AC noise to. You should never connect it to NEG terminal.

[@iceberg789]

whoever comes up with such bad design ideas.

[@guywilkinson]

Funny how you blame the account managers rather than the engineers.... maybe you bad experience in your engineering career. For me it's usually both or a New Product Introduction team 😂

[@demef758]

Your 2-minute rant about some imagined penny-saving design house is absurd. Yes, the design is flawed, but it was simple ignorance of grounding principles that was root cause. For most EEs, proper grounding is black magic, as is magnetics. These things are almost always learned from experience.I bet that no more than 1% of the commenters here understand grounding principles. Here's a simple test for everyone: "what is 'ground'?" Better yet, "what causes noise on a ground plane?" If you can't answer these questions, should you be banned from commenting here? This Chinese design house screwed up their grounding on a $50 power supply, and you act as if he's now a millionaire now because of it. Come on, man, you're smarter than that!

[@BlondieSL]

Yeah, that rant almost have me X out. Just a worthless waste of time.

[@k-tech796]

Obviously you never had argue in hour long meetings for why you need that one 0,04 Cents more expensive flat head screw. Most of it was meant to be funny, but there is a kernel of truth.

[@BoxingDayAC]

K-Tech is for the children

[@tactileslut]

Good thing I prefer my supplies and measurement tools to float. Unexpected ground loops can be confusing or dangerous.