Core Saturation Measurements In 70.7 Volt Line Transformers Used For SE or PP Tube Output Stage

Edit:
It looks like after much back-and-forth and further research, this method is very flawed and the results are incorrect. I sometimes leave flawed videos posted as it documents what does not work so others don't have to try it. I am working with other viewers to find a better method.
........After many measurements on the small 70.7 volt line transformers where inductance and impedance values were turning out so good, I made a measurement of inductance with DC current flowing in the primary and found that even just a few mA (7 mA or less) would reduce the primary inductance from 6H to 20 mH or basically no primary inductance at all. The conclusion is these transformers can not be used directly in a single ended (SE) vacuum tube amplifier. There may be another method of shunt feeding the output tube with a second inductor and isolating the 70.7 volt line transformer with a capacitor that might work. I doubt these transformers can be used in push-pull amplifiers either as any plate current imbalance in the push-pull stage will likely saturate the core of these transformers.

[@owenlaprath4135]

Don't panic! They work fine even for single ended. I am using them for SE with a small modification, that just takes a little patience. I actually opened the wrap-around clamp on a pair and took a rotary tool with a cutoff disc to cut a gap into the core, and the pair works dandy for a single ended stereo amp with EL84 tubes. Here you go, all the taps on a 10W or smaller line match transformer. First column is the markings, second column is the resulting impedance, third column is what to connect. BEWARE that there are some of these, where the wire gets thinner on the higher impedance side. Use only those for tube output, that have constant wire thickness! 70V lin-match Rpri(Ohm) = (70V^2)/Ptap Uin = 70V Rsec = 8 Ohm Usec = sqrt(P*Rsec) = sqrt (P*8) Ptap Rpri Rpri plate-plate Usec Uin/Usec Rin/Rsec (Uin/Usec)^2 0 0 plate-plate 4K,8K 10W 490 8.94 7.82 61.15 5W 980 B+ 4K 6.32 11.07 122.5 2.5W 1960 B+ 8K 4.47 16.65 277.2 1.25W 3920 plate-plate 4K 3.16 22.15 490.6 0.62W 7840 plate-plate 8K 2.23 31.39 985.3

[@ElPasoTubeAmps]

I think you have done an excellent job of helping others understand this. For more than 50 years I have used transformers without a gap and run current thru the high Z winding and modulate class-C amplifiers for AM. Some modulation transformers are specifically labeled to handle current thru the secondary but in my case, there were times I only had-what-I-had and used it and it worked just fine, i.e. non-gaped transformers with current thru the primary or secondary. There is one exception I had and that was when I build a SE 833A amplifier and used a KW level output transformers designed for push-pull 810's and it performed poorly at low frequencies. I think it was OK down to about 80 Hz but below that it was poor for LF response. Further investigation showed me that the quiescent current of the 833A was about 150 mA and at that constant DC current the primary inductance dropped to a few mH. As for using these same transformers running current thru the primary that would normally saturate the core, I was only running audio voice modulation for an AM transmitter where frequencies of the voice rarely dip below about 300 Hz and they worked just find. I have posted videos of it. Thanks for your comments and the time it took to document the information. Stay safe...

[@ot4kon]

Where do you cut and how big the gap was?

[@aholatom]

Yes the transformers intended to have a single ended anode current flowing through should have a air gap in the core to prevent saturation. It would be more clever to have the secondary wound around the choke instead of having both a choke and transformer in the design adding weight and cost. The coke core must anyway be designed with an air gap to handle DC. I'm currently repairing an old Philips radio where the audio is strongly distorted. I'm starting to suspect there might be a saturation issue in the transformer. It is a single ended one-tube output stage where 56mA anode current goes through the output transformer 3.9 H winding. Center tap is to 220V anode voltage. There is a 1 kohm 4W power resistor from the opposite 13 mH winding from the center tap. This resistor feeds 40mA DC current to other tubes anode voltage that is 180V, filtered by 50uF capacitor. This invention seems to intend to cancel the magnetic field in the core by the opposite magnetizing DC current. However, it does not work as the amount of turns is much less to the compensation winding, or the compensation current is too small. Maybe the transformer has a short in the compensation coil. How to measure shorts in a transformer when you don't have the specification of the transformer?

[@ElPasoTubeAmps]

That is an interesting design. Either it is designed to attempt to cancel the field in half the core or use it as a power supply choke. In any case, I think I might rearrange it. As for testing it, if it is indeed a CT winding, the DC resistance should be close from CT to each side and, of course, not shorted to the core or the secondary. I am sure you want it to work as-designed. I have started using a different method to measure inductance with a DC bias thru the inductor. Seems to be very good. www.vacuum-tube.eu/wp/wp-content/uploads/articles/Measuring_Unknown_Choke.pdf

[@EJP286CRSKW]

Even if you can balance the DC perfectly, or get rid of it via series C, a few mA across 8K is still only a fraction of a watt you can develop before saturation with AC at low frequencies.

[@MarkTillotson]

Your idea of adding a DC bias while measuring the inductance only works if the DC bias supply is constant current, ie has a far higher impedance than the primary winding itself at the test frequency. Putting 700 ohms across the primary basically turns it into a 700 ohm resistor with a tiny tiny bit of reactance - paralleling a small R and large L completely bypasses the impedance of the inductor. Adding the DC bias to the secondary has the same problem since the secondary load/source impedance is reflected in the primary. So a constant current source into the secondary would be needed too. Another possible way to add the DC bias is through a series choke (which has much larger inductance than the winding it feeds). So either test with a constant current source to add the bias, or measure the saturation by looking at the response on a curve tracer to see the AC current that starts to saturate the core, since that's the same phenomenon basically (this basically measures the max usable power at the test frequency). Once you know the saturation current for AC, your max DC current is going to be some fraction of that (you need to leave headroom for the actual AC signal current - I'd suggest DC offset about 1/3rd of saturation limit is a reasonable compromise). I guess cutting a thin enough slot in the core isn't a viable option without access to fancy tooling?

[@ElPasoTubeAmps]

I am hearing you and I agree with you. I recently found a procedure that describes much of what you are saying. Add a large L in series with the DC to basically keep the AC out of the DC supply and a capacitor in series with the AC to keep the DC out of the AC supply. Here is a link that I think is going to work. Now all I have to do is figure out how to make it real in my shop... www.slideshare.net/vishalgohel12195/owens-bridge-and-measurement-of-increment-inductance I really appreciate your thoughts and suggestions on this as it is right in line with what I am beginning to understand. As for cutting slots in the core to reduce saturation, I definitely don't have the tools for that and I wouldn't want to anyway as I want to discover and document the parameters of these little transformers as they are.

[@josepheccles9341]

I was lucky enough to get good results from using a pair in push pull configuration with EL84 tubes. It is a very low power system I built, but sounds amazing. The ones I used are rated 16 watts line matching that I got on ebay for about 7 dollars each . And the tubes are a matched quad I got on ebay for about 12 dollars. They are pretty weak. The unit makes about 1.5 watts per channel.

[@ElPasoTubeAmps]

Joseph, very nice you built with them and shared your results. I want to do all this also but there is just so much time and so many projects.. I suppose that is a good thing. I am going to guess you used the 8K or 16K winding for your primary? Are you using any NFB? I can see about 1.5 watts with no NFB but with maybe , 16 dB of NFB it might rise to 10 watts or so. I am in the process of building up another amp using a Chicago Standard BO-13 OPT. It measures 40H+ on the primary. Most all 20 watt level OPT measure 20H so I am expecting to get really good LF response. It will be interesting to see how the HF performs. Thanks for sharing and your comments. Stay safe...

[@josepheccles9341]

@@ElPasoTubeAmpsI was using my phone as a source, so I don't know exactly how the roll off would look. I really don't have any test instruments, So I just plunk around until I get a sound I like.

[@Claude77]

You could use a current source in place of the plate choke, a current source is easy to build, a high voltage transistor and a couple of diodes :) should work really nice :)

[@PeterMilanovski]

I have been looking for a follow-up video on this topic... Did you try the circuit with the DC choke bypass? I want to try this with a solid state amp but I don't have a line transformer to try it yet. It's definitely on my high priority list of things to buy.... We have a store here in Australia called Altronics that has a good range of transformers in many different power capacity options, though one thing that might be of interest to you is their DIY vacuum tube amplifier kit. I believe that it was designed by Silicon Chip Magazine and they have come up with some really good stuff over the years... Well this particular tube amplifier kit is using rather small output transformers, I initially thought that it wasn't using any output transformers because I couldn't see them hiding behind the output tubes LoL... So it has me thinking about why and how they are getting away with using such small output transformers on a 20W tube amplifier! At $650Au it puts it out of my price range just to find out. I have a vintage solid state amp which was also designed by Silicon Chip Magazine and it's an amazing amplifier so I have no doubt that this tube amplifier kit will be no different... There will be an article on this design in silicon chip which there must be a mention about the output transformer choice and I'm on a mission to find out more about it, so yeah! I thought that you might be interested in this modern day rendition... As for the line transformers, I would love to see and if possible hear what those little things are capable of...

[@Mcandmar]

The older versions of the Bottlehead S.E.X single ended amp kits used those transformers with a choke plate load and capacitor coupled output as per your diagram. Problem was those are original Speco 7010's, the later ones are physically much smaller and had different specifications. Be warned if you are thinking of buying a pair, the current production T7010's are not the same transformer.

[@billmoran3812]

I think it would be interesting to see what the effect of the B+ inductor and capacitor would have on the overall frequency response of an amplifier. I'm thinking the high end would roll off dramatically. Might be fun to see. Wonder about parasitics too. Hmm. Might be a reason why it isn't done.

[@ElPasoTubeAmps]

I forgot to remove the screen connection in the model from the UL tap on the transformer. It does work but produces about 1/3 the output (so far). Needs some more research and modeling of L and C values. It looks like it might work reasonably well with the best values but reduced power.

[@moodyga40]

this works very well in push pull

[@maxjurgen1900]

But we can disassemble the transformer and than assemble again as SE transformer with gap.

[@ElPasoTubeAmps]

I did the best I could at the time I made this video but I didn't think to address core saturation so I don't actually know if these transformers have a gap for SE triode operation. I tend to think they do not have a gap as they are used between the output winding of an amplifier and a speaker system and have no DC component in them. I need to look into these further but i am convinced they could be used in PP circuits.

[@grantwills5421]

Thanks for your work in testing these transformers. I have a special interest in your tests as I use numbers of different PA transformers successfully in my valve amp designs at www.valveheaven.com I decided to replicate your tests and confirmed the apparent dramatic dropoff in primary inductance with a PA transformer similar to the one you tested. To investigate this further, I tested a large quality gapped SE output transformer and found that it too appeared to demonstrate the same dramatic dropoff in inductance with only a few mA of DC current flowing in the primary. This confirmed for me that this method of testing primary inductance with DC applied must have a flaw that I can't quite identify. I think it has to do with using coupling caps to isolate DC from the inductance meter. I then decided to to take a different tack with this test - to connect the inductance meter directly to the primary and to apply the DC current instead to the 8 ohm secondary of the transformer. With a 5 ohm series resistor connected to my bench supply, I was able to apply a variable DC current up to several amps to provide a significant magnetic offset in the core. I was not surprised to see only a small decrease in primary inductance even with amps of current flowing in the transformer secondary. So all this to say that you can be reassured that these transformers work very well in valve output applications - I use them successfully in amp designs from 2W to 40W output. Cheers

[@ElPasoTubeAmps]

I really appreciate you posting this. I did not try it your way by putting current through the secondary and if you did not get the inductance dropping to virtually zero as the core is being magnetized as I did then it does seem there may be a flaw in my test procedure. I will revisit my approach. In the GR manuals they speak of incremental-inductance measurements where they do carefully increase current through the windings of a transformer/choke and measure resulting inductance. That is the idea that I based my measurements on. I will try to duplicate your measurements and see what I get.

[@ElPasoTubeAmps]

Grant, Here is a procedure that will probably work and what I will attempt tonight, time permitting. I have to think this thru carefully so I don't so something really dumb and damage my GR instrument. Take a look on page 9 and 10 figure 5 and 6. I like the way you made your measurements of magnetizing the core by applying current thru the secondary but I am not sure that is valid. I will also measure some much larger transformers, as you did, and see if they saturate at just a few mA. If they do, then there probably is something wrong with my method. Here is a link to the GR inductance bridge and the instructions I mention above. www.ietlabs.com/pdf/Manuals/GR/1632%20Inductance%20Bridge.pdf

[@ElPasoTubeAmps]

I did do some quick measurements and with the GR recommended series method, I get 4.37H with 11.3mA flowing through the transformer primary and 3.28H with 18mA flowing through the primary. It looks like my capacitor isolating method does not work. Hmmm... it is interesting that I do get the same inductance values when this is used with a standalone inductor and does not have a bridge connected. I wonder why. In any case, now it does appear that the L of the line transformers does not drop to an unusable value. Certainly builders that have used these transformers report good results. It is nice to see (and hopefully it is true) that the core does not saturate as quickly as I thought. Looks like this is a good project for some in depth measurements.

[@trh4246]

Thank you for all your great videos-- lots of fun to watch. Could you please explain the electronics of why the tapped winding, which seem to be effectively two 1.5 H inductors in series results in a 6H total measurement. I'm guessing it has to do with the coupling of the two windings through the core but I haven't been able to find any info on the web that explains it. Thanks very much! Terry

[@ElPasoTubeAmps]

It does have to do with the coupling. Assume you measure the inductance (L1) from CT to one plate lead and then measure the CT to the other plate lead (L2) and find that they are equal (within reason) then you take the square root, in this case of 1.5H and get 1.22. Then you take the square root of the other side, and in this case they are equal and take the square root of that 1.5H and also get 1.22. Then you add the two sqrt roots together (1.22 + 1.22) = 2.44. You then square the 2.44 and get 6H. That is how they combine. In equation form it would look like this: [sqrt(L1)+sqrt(L2)]^2 Hope this makes sense.

[@trh4246]

David, Thanks for the reply. I'm still left with my original question, which I didn't state very well. I fully understand the equation-- [sqrt(L1)+sqrt(L2)]^2. My question, I guess is: where does the equation come from? Two inductances in series normally just add. What in transformer theory leads to this more complicated equation. I will freely admit my lack of transformer theory :)

[@ElPasoTubeAmps]

Yes, I understand what you mean now and inductors with no magnetic coupling can be treated like resistors. I have the same lack of transformer theory so, I don't know either. :-) A few months back a good gentleman and I got into questioning where did the fact that we have to use radial frequency (2*Pi*F) for reactance values come from? What's wrong with just cycles per second? Why must it be radians per second? (maybe you know..) We beat it to death going to dimensional analysis and everywhere in between and never really got to a satisfactory answer. I think, first of all, to really understand anything we have to go head-first into the discipline, like transformers, and develop the advanced terminology and language of that specialty. I am interested and I intend (and hope to succeed) in accurately and consistently being able to measure incremental inductance in transformers and chokes with a jig I can make and procedure someone else came up with . I finally understand, so far, it takes a DC and AC component to measure incremental inductance and not just a DC core magnetizing force. It was a challenge to be able to be able to measure high inductance values in iron-core inductors/transformers until I finally came across the procedure shown in one of my videos. The math is not staggering but it is not high school either - at least not at my high school of the 1960's.. I figure the learning curve for measuring incremental inductance is going to be an equal challenge. Maybe then I will understand the reasons for the necessary formula(s). Then again, I sometimes wonder if the guys that came up with these types of formulas didn't derive some from reverse engineering of empirical measurements. I can just see the guys 100 years ago measuring inductance with their apparatus. One guy measures one half of the transformer and then the other half and then the whole thing and says, "what's wrong with this picture?". All the while another fellow is over there with paper and pencil and his slide rule (invented in 1622) and he says, "hey if you take the square root of this one and the square root of the other one and add them together and then square the sum, you get the right answer. Of course, it probably took him two hours to do that and the first guy looked over at him and said, "smart-ass, let's go have a beer". And thus the formula was born. That's how I think it happened. :-)

[@ElPasoTubeAmps]

Here is the real answer. I am glad you guys ask these questions as it makes me do research. You can jump from one tutorial to another at the top or the bottom of each one. Very nice article. www.electronics-tutorials.ws/inductor/mutual-inductance.html The answer starts with section (3) Mutual Inductance as M^2=L1*L2 or M=sqrt(L1*L2) Then in section (4) Cumulatively Coupled Series Inductors, we get, Ltotal=L1+L2 +2M so that Ltotal=L1+L2+2*sqrt(L1*L2) This is the formula and is equal to the one we already discussed.

[@trh4246]

David, Thank you very much for researching this and finding those great tutorials. I've done a quick scan of the material and see that it will take a deeper read to take it all in. Looks like an opportunity to sit down with a cup of coffee (perhaps more likely a beer) and learn something new. Terry

[@PaulinTaegu]

PP works fine ElPaso. Try it with 6AQ5s at a B+ of +250VDC. I think you might change your mind about the PP application. ;^)

[@ElPasoTubeAmps]

I am sitting here think about this. Your experience says it works and there is a link that shows where a fellow has also had real success. But my inductance values are so low with tiny amounts of primary current. I don't get it. I am attempting to model the circuit with the shunt inductor for the plate load and it is turning out to be a disaster.

[@frankgeeraerts6243]

Curious about the plate load inductor & capacitor approach .....to gain something often one loses something ........finally whatever the approach , it's the result that counts . They have nice plate chokes at Lundahl. vinylsavor.blogspot.be/2011/12/new-plate-chokes-from-lundahl.html www.jacmusic.com/lundahl/html/plate.htm There's also the solution of John Broskie .......

[@AC9VC]

DC current flowing from the center tap to both ends in a PP will not cause the core to saturate like an SE setup because equal current is flowing in both directions around the core and they cancel each other out.

[@PaulinTaegu]

For sure in pure class A operation, where both tubes are on continuously. AB1 has an output tube shutting down part of the time but it still seems to sound fine for my little guitar amp applications! ;^) Little low wattage SE OTs from old radios, etc. seem to be more common and cheaper. Finding a cheap little OT for lower wattage PP applications was a problem for me before I stumbled onto this solution. 8^)

[@ElPasoTubeAmps]

No doubt we all agree on that but I am concerned that even a couple mA of imbalance might tend to saturate the core on these little transformers. It might be required to put a cathode current balance in and keep the current low. Some people report good performance with fairly powerful output tubes, i.e. 6AQ5, 6V6, etc. I think I would keep the output tube to something like a PP 6SN7 so that the current is naturally low (a few mA) and maybe ground the cathodes (cathode bias) with a pot with the wiper to ground so that the currents through each tube could be carefully balanced. From what I have seen so far, it seems the Achilles Heel of these line transformers is any amount of DC through the primary.

[@Beelzybud]

In the 'Choke-Capacitor-PP Output Transformer' scenario; would a resistor work as a substitute for the choke? Not asking if it would be a good idea necessarily, just, would it work. :)

[@ElPasoTubeAmps]

Yes, a plate load resistor would work. The idea of using the choke is to provide more gain and drive power to the final amplifier stage. This would be necessary if the output is a large power triode that needs the grid driven into conduction. In any smaller case, like driving a 6V6, this higher gain in the driver would not be necessary. It is funny how we get into some of these ideas for audio amplifiers. In an attempt to use a small output transformer (note the word "small") we start wanting to use heavy-iron like 250-400 H chokes so we can build a 75 pound 10 watt per channel amplifier... Who knows what gremlins are going to exist in such an amplifier with a plate choke connected thru a capacitor driving current into the grid of a large triode. It's just crazy. With all that said, I have never build such and amplifier although I believe there is a guitar amp made by Fender, way-back-when, that does use a plate choke in a low power amplifier. In the creation of music I can see using such a circuit as it may just provide the sound the musician is looking for but for a play-back system, it seems pretty outrageous to me. I have made a couple of videos on phase inversion for PP amplifiers and one example was a CT choke. It will work but I never built it and tested it and lived with it. Also I made a short video on the phase shift of using transformers vs RC coupling in voltage amplifier stages. The idea that some seem to believe in that everything was better when amplifiers were made of lots of iron and everything has degraded since we left SE amplifiers with no NFB are still alive and well. Personally, I think the golden era of tube amplifiers was from the late 1950's to probably the late 1960's. There is little to no value or gain in "designing" or "redesigning" tube amplifiers. It's already been done and tested and sold decades ago. Pretty much all we need to do in today's world is duplicate known high quality designs. The three best known designs are the McIntosh, Dynaco, and Williamson (for PP amps). I have built a good bit of very high power amplifiers (RF) and these devices get really Hot, so, when I see these amplifiers playing on KZbin with 100TH, etc, I know these devices heat up the chassis beyond what anyone would expect unless they have had experience with these tubes - and that doesn't even touch (no pun intended) the fact that the exposed plate connection of these tubes run at KV levels and could kill anything that touches them. Just my soap-box speech. Thanks for your comments.

[@bundylovess]

Nice one big thumbs up 👍

[@bobh.3705]

So now you have me wondering. Is a DC balance pot needed on a PP stage? I think the Williamson design has one. Not to over burden you with work but what is the typical DC unbalance on these PP stages? And of course it begs the question, how much DC current will a PP transformer like the Acrosound take before it affects operation?

[@ElPasoTubeAmps]

I am modeling the choke fed driver stage right now and things are turning out very well. I am surprised and pleased. Some transformers specify maximum DC imbalance. I would think in the case of these little transformers the DC balance/ imbalance current would be critical. As for your question on a DC balance pot in a PP stage, in my opinion, it is necessary and I know for a fact that balancing the cathode currents in the output stage makes a significant improvement in performance (lower THD). It is also a very good idea to balance the drive (signal) voltage into the grids of the output tubes. An excellent example to study that has both of these adjustments is the McIntosh MA230 amplifier stage. The schematic is easy to find on the Internet. I have implemented these static (cathode current) balance and dynamic (AC grid signal) adjustments into the original design of the Dynaco Mark III amplifier with significant improvements in lower THD. I am very much a believer in balancing the currents in the power amplifier output tubes and the drive voltage to the output tubes.

[@ElPasoTubeAmps]

I think those are great and fundamental questions. The Williamson does provide DC cathode balance in a (cathode bias) circuit. I think you are going to use fixed bias (?) so, the little circuit used by McIntosh in their MA230 amp circuit is made to vary the bias voltage and balance the cathode currents. They also use a pot in the phase inverting stage to balance the dynamic (signal) drive to the finals. Both of these methods lower THD and aren't too hard to implement. I know some transformers specify maximum DC imbalance and I sure don't know about the Acrosound or any others but if the balance circuits are implemented, that should take care of any concern. I am working with the model for the shunt fed driver stage. Turning out great. Better than I ever expected. Seems all is good from 10H up whether the inductors are linked (like a CT primary winding) or separate chokes. It appears the coupling capacitors work well at 100uF. Low values like 5uF seem to do funny things to the output current.

[@vidasvv]

Another great video ! 73 N8AUM

[@cerrem1]

Nice video.... Your inductance measurements are not done right....therefore the inductance number are way off..

[@ElPasoTubeAmps]

I agree. I think I made some major errors in making these measurements. Maybe I should try again... :-)

[@W1RMD]

Thank you for the video. I've built a 6L6 amp with the bigger brother transformer to this one. It's an Episode SPKT-25 25 watt transformer with the lowest wattage winding being 5 watts (Ebay). I get HORRIBLE distortion below 93 hz! A huge disappointment when I've just spent $185 in other parts, only to have to buy a $$$ Classic tone $$$ transformer anyway. Do yourself a favor, BUY THE GOOD TRANSFORMER!!!! You'll be sorry if you don't! I'm curious if anyone here has successfully built an amp using one of the 70 volt line transformers. Another thing to note, the dc resistance of the primary is only 30 ohms. A "good" audio output transformer measures about 300 ohms. I wish I could have read this comment I'm writing now BEFORE I bought this transformer!