Investigating THD and Frequency Response on Coupling Capacitors in Vacuum Tube Amplifiers

แชร์
ฝัง
  • เผยแพร่เมื่อ 8 พ.ย. 2024

ความคิดเห็น • 34

  • @Betterhifi
    @Betterhifi 8 ปีที่แล้ว +2

    Leaky caps are usually "noisy" caps. It looks like a lot of the THD was actually intermittent noise. It's also not hard to believe (It's expected!) that any well designed triode amp would sound much better than a Mcintosh amp. McIntosh amps are designed to "measure" well, not "sound" well. Too much feedback, both globally and locally. The Dynaco's usually sound much better while stock and dramatically better after some good mods! A well designed triode amp would be better yet! Great video!

  • @PrinceWesterburg
    @PrinceWesterburg 3 ปีที่แล้ว +1

    My valve hifi has so much noise on it, I've been putting off replacing caps but now I'm ON IT!!!! :D

  • @brittisaacrossman
    @brittisaacrossman 9 ปีที่แล้ว

    This was really helpful! I've been debugging a problem amp for weeks now, tearing my hair out trying to find the issues... I dont have an ESR meter, or anything to measure caps other than my DMM with cap setting (which works for most coupling caps). Other than how much they've drifted, I haven't been able to tell much about what's happening, other than there is incredibly high voltage all over the place (+20% of spec everywhere). I think the amp is seeing a bad load, which could be caused by poor biasing of the power amp for the reasons you showed in your video. Then again, I'm no EE...

  • @TheFRiNgEguitars
    @TheFRiNgEguitars 7 ปีที่แล้ว

    Thank you for this informative video. I'd throw out the leaky cap anyway, even if used in a tone stack (where it would not leak DC) for the concern its value may change radically and bypass more resistance and change the amplifier's tone contour.

  • @benadams6332
    @benadams6332 9 ปีที่แล้ว +1

    Thanks for sharing your experiments and findings.
    Would be interesting to see a plot from an Impedance Analyzer frequency sweep using that capacitor compared to a new one. Mostly parameters |Z|, Rs and D.

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  9 ปีที่แล้ว +1

      +Alfred Stampe I did that tonight and could not see any difference at all. I was a bit surprised. However, when I put the equivalent shunt resistance of 8 meghoms across the good capacitor in the amplifier, I got the same bizarre THD response as I did with the bad capacitor. Watching the oscilloscope, I could see about a 1 Hz oscillating bounce in the amplitude of the signal. Possibly something unique to my amplifier circuit. i will try to make a video of all of this.

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  9 ปีที่แล้ว +2

      +Alfred Stampe i might add that all i really got on the sweep is an exponential curve that one would expect from increasing frequency which we might call |Z|. As for Rs that would have to be derived from D which was about 0.01 on the bad capacitor and as close to zero as I can read on the GR for a good capacitor. I also constructed a circuit with large NP capacitors, to protect my equipment from the HV DC, and then placed 400 VDC across the bad capacitor and re-scanned thinking that the DC bias (stress) on the bad capacitor would make it act crazy and got nothing different - the plot was the same as without DC across the capacitor. I believe it is possible to completely evaluate a small value capacitor but it is time consuming and takes some involved measurements - probably easiest to just replace it if in doubt.

    • @fpliuzzi
      @fpliuzzi 9 ปีที่แล้ว

      +ElPaso TubeAmps In some future video could you please elaborate on your method of applying a DC bias voltage to a capacitor while testing it which both protects the measurement device and does not interfere with the accuracy of measurement result. Many of us with LCR meters (such as the DER EE DE-5000, etc), would like to do measurements of inductors/chokes and capacitors whose values can sometimes be highly voltage dependent (i.e., some multilayer ceramic capacitors). Thanks again.

    • @gman76utube
      @gman76utube 8 ปีที่แล้ว

      +ElPaso TubeAmps I looked at the TO-250 schematic and see that the 1st and 2nd stages use 100k plate resistors, meaning that the tube ought to be biased at 2mA in order to drop half the supply voltage, giving the plate voltage equal room for positive/negative signals. By introducing a leaky cap, you change the bias points of the 2nd stage (and possibly the 1st stage). You tested the bad cap at 400V, but the actual circuit applies only 200v across the cap, therefore probably not 50uA of leakage, but some nonetheless. You measured the grid-cathode of the 2nd stage to change from 7.3v to 4.1v. This changes the bias current from 2mA down to 1.14mA, pushing the plate voltage higher. 400v - 100k*1.14mA = 286V. This is likely the cause of the high THD readings. The 2nd stage is operating at a much different point in the transfer curve of the 6SN7.
      I suggest probing the plate of the 2nd stage, also check the plate of the 1st stage. It would be nice to know the grid, cathode, plate voltages of all the nodes.
      By the way, to calculate the bias of the 2nd stage, I see that the tubes share a common 1.8k cathode resistor. I assumed half the impedance (3.6k) to derive the plate current of a single tube.
      As sporadic-Z pointed out, you should report the grid voltage (negative) or grid-to-cathode instead of reporting the cathode-to-grid (positive). Triode datasheets show plate characteristics with various Vgk voltages. I am curious if the grid of the 2nd stage (normally at 0.0V) was pushed up above GND due to the leaky cap.

  • @MrDeeejaaay
    @MrDeeejaaay 9 ปีที่แล้ว

    Thanks for all you Interesting and informative videos. I check every night to see if you have something new out.
    I understand your thought process on the possibility of a bad cap not causing any serious problems if it's upstream of the tone stack but wouldn't the THD still be affected? It may still cause a problem, but it will be that much more difficult to track down. Just a thought. - Doug.

  • @59seank
    @59seank 9 ปีที่แล้ว

    I love your experiments. Thanks for sharing.

  • @jjswan1953
    @jjswan1953 5 ปีที่แล้ว

    Thanks, the only I could find spelling out the details of caps and thd

  • @EJP286CRSKW
    @EJP286CRSKW 9 ปีที่แล้ว

    Changing the bias point by nearly 50% +ve must inevitably have a huge effect on THD. You would run into grid current and saturation much sooner. But I think the readings you got were anomalous. It would have been interesting to see the THD residual on a scope via the AA501.

  • @fpliuzzi
    @fpliuzzi 9 ปีที่แล้ว

    Very interesting. Could you temporarily shunt 8 Megs of resistance across the new, good capacitor to simulate the leakage of the bad capacitor? It would be intriguing to see if the T.H.D. measurement would be as poor as it was when the bad capacitor was installed, or if an additional defect in the bad capacitor was skewing the results somehow.

    • @fpliuzzi
      @fpliuzzi 9 ปีที่แล้ว

      +fpliuzzi Additionally, I'd expect the two distortion measurement results to be similar, but I was curious if the T.H.D plot would be as "lumpy" as it was when the bad capacitor itself was installed.

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  9 ปีที่แล้ว

      +fpliuzzi That is a good thought. I wish I had done it with the good capacitor. I will try it and if I don't make a new video, I will post the results in a "speech bubble" in this video. I think a steady resistance of 8 meghoms will make it behave very differently than the bad capacitor. In any case, it seems the results from testing the capacitor for DC leakage is definitive.

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  9 ปีที่แล้ว

      +fpliuzzi Just tonight I put 8 meghoms shunt across the good capacitor and got the same bizarre THD results. I didn't expect that as I figured the bad capacitor would be a non-linear resistance to frequency. It also induced a slow, about 1 Hz, bounce in the amplitude of the output of the amplifier. I am convinced it would have somewhat specific results in different amplifiers but in all cases, it seems THD will go very bad with additional anomalies.

    • @fpliuzzi
      @fpliuzzi 9 ปีที่แล้ว

      +ElPaso TubeAmps Very interesting. Thanks for the additional test.

  • @Dan79istheman
    @Dan79istheman 9 ปีที่แล้ว

    Thanks heaps for the insight David

  • @MiddletonGuitars1
    @MiddletonGuitars1 9 ปีที่แล้ว

    I know you mentioned that you could use the amps' power supply if you don't have a separate one to test the capacitors at HV: would it work and possibly save the step of desoldering parts to test a coupling capacitor in-circuit and put the ammeter from the output-side of the cap and to ground?

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  9 ปีที่แล้ว

      +Nick Middleton l think you are going to have to remove the capacitor and test it independently of the circuit.

  • @jimcatanzaro7808
    @jimcatanzaro7808 2 ปีที่แล้ว

    So I’m going to work on changing the tone capacitors how can I tell the positive and negative on the new style poly tone capacitors I was reading and I was told there is a positive and negative side are they right and if so how can I check them

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  2 ปีที่แล้ว

      I think this link will answer your question.
      www.aikenamps.com/index.php/where-to-connect-the-outside-foil-on-capacitors
      I have not verified this but it might help.
      th-cam.com/video/tFa6XJccvTE/w-d-xo.html

  • @lostigresdelnorte
    @lostigresdelnorte 9 ปีที่แล้ว

    Will there be any difference in audio quality?

  • @jimcatanzaro7808
    @jimcatanzaro7808 2 ปีที่แล้ว

    I have a few 1962 Hh Scott 222c amplifiers do you think I should replace all the coupling capacitors or leave the org ones in I don’t have all the testing equipment like you but it runs and sounds good but not as great as I would like it ?

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  2 ปีที่แล้ว +1

      If it works and sounds good and the tubes are not screaming hot, I would be careful about working on it. It can be perfectly OK .
      I agree that with a lab full of equipment, it is easy to assess things like capacitors. On the other hand, it is actually simple to check if capacitors are "leaking DC". Of course, the capacitors most likely needs to have at least one end removed from the circuit so it is "floating" and the capacitor can be tested independently of the circuit it is in. Putting a power supply in series with the capacitor and a volt meter is really all it takes and if the volt meter doesn't eventually go to zero volts, as the capacitor charges, it will indicate that the capacitor has DC passing thru it and that will seriously affect the circuit the capacitor is driving. A variable high voltage DC supply is necessary and you don't want to put a voltage across the capacitor and volt meter greater than the voltage rating of the capacitor. For a short time a little over the rating of the capacitor would be OK but not for permanently.
      I have "old" capacitors that are as good as anything new but there is a mentality on the Internet that electrolytic and coupling capacitors should be replaced in all vintage equipment. If I were a commercial repair shop, that might be a good philosophy for the customer not just to sell more capacitors and service. On the other hand, I have worked on a good bit of equipment and for some reason I remember the Fisher amplifiers having bad coupling capacitors every time I serviced them. Maybe they ran them too close to their rated voltage which does seem to shorten the life of capacitors.
      I am sure there are still a few service providers out there that are honest and can provide good repair service but I would not know where to find them. We are kind of like the toaster repairmen - not many left... lol
      If asked, I tell people that are buying vintage amateur radio equipment like Collins Radio, for example, to be careful when purchasing such beautiful instruments because you are probably going to have to repair the equipment yourself when it needs it.
      One last thing in my babbling on, and that is much of the older equipment was made to be repaired. All equipment needs repair eventually and the older equipment builders knew this and there were zillions of parts made "back then" and they are still around and new parts made also. A very different philosophy from todays equipment when it breaks, we toss it and get another one, fancier and prettier than the last one. Hope this helps.

    • @jimcatanzaro7808
      @jimcatanzaro7808 2 ปีที่แล้ว

      @@ElPasoTubeAmps well I have 2 of these HH Scott 222c amplifiers so maybe I will change the coupling caps on one amplifier
      I have already installed new tubes and filter cans and tested them on both amps one amp the preamp sensitivity is very different with bass and treble. I also dropped the voltage down since my mains is 125v the B+ was dropped to diagram specs on both and changed the rectifier to silicon for the negative bias and I had to Drop that to spec
      But I would love to see a test with your equipment and test it you run the neg bias higher or lower then spec and see what the output values would be that would be interesting I’m sure there are videos made already
      Thank you for your videos they are inspiring and keep me in love with audio I will have to watch a few videos on how to install the caps on the amplifier I’m very critical on wire placement and long term relatability I do see about 10 caps they look like wax sticks and some look like they got hot and deformed the value is .25 mf

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  2 ปีที่แล้ว +1

      ​@@jimcatanzaro7808 I see you can do what is necessary to keep it running properly. I am a fan of the Orange Drop capacitors and don't spend money on the fancy high $$$ capacitors. I assure you we can not measure any difference in performance with test equipment by using these huge and expensive capacitors but I can never say how it will make an amplifier "sound" as that is purely subjective.
      I did a video a few years back by making the bias on a Fisher more negative, lowering the temperature of the 7591's and the measurements were significantly poorer with more negative bias. More negative bias and less plate current pushes the tube closer to crossover distortion and raises THD. it is an unfortunate fact that the hotter we run our output tubes, the better they perform. We get more power and less distortion by running them "hot" right at max plate dissipation. Sad but true. So, what is the objective ? I suppose a compromise and just admit we need to replace tubes occasionally or if we keep the power down to a few watts, we can run our tubes a little "colder" and maybe lengthen the life of the entire instrument.
      Power and THD are obviously not the final work in an amplifier. It that were true there would be no such thing nowadays with SET amplifiers. The harmonic profile of a SET amplifier approaches a triangular wave with every harmonic from the second upward in an ever decreasing amplitude - but somehow we can love the sound of the SET amplifier. Just think back to the days of the audio engineers whose objective was to get rid of that horrible second harmonic that dominates single ended amplifiers and they did it with the push pull amplifier. Also, the wanted to make a 10 watt amplifier weigh less than 60 pounds.... now we build 60 pound 7 watt amplifiers and think they are the greatest thing ever. We have to laugh at ourselves a little bit. Stay safe...

  • @meshplates
    @meshplates 6 ปีที่แล้ว

    I believe you could achieve superior distortion figures by reconfiguring your driver stage. 6B4GS are more demanding of the driver stage than EL34s. Look at the ALTEC A 100 A circuit for a good example of choke loaded cathode drive.

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  6 ปีที่แล้ว +1

      It is interesting that you bring up the choke loaded cathode drive as that is what I am considering building into a PP SV572-10 amplifier and pretty much copy the Altec 1570B circuit. I am sure I will use 6V6 instead of the 6K6 although I have a number of 6K6's that I can play with. I plan on using one of the line transformers I have made videos on recently, using only the primary of course where I can derive a CT.
      As for the 6B4G, the bias is very negative, if I remember correctly around -50V or so, and at the level I use it (below 10 watts) I don't think I am driving the grid positive at any point. In general I copied the Acrosound schematic for the 6B4G amplifier.
      I looked up the Altec A-100 and found the manual for the organ but not the amplifier. If you have a link to the schematic, I would appreciate it if you could forward it to me.

    • @meshplates
      @meshplates 6 ปีที่แล้ว

      ElPaso TubeAmps sorry i didn't see your answer. The best link is this one but you have to scroll down for the Altec A100A. This is one awesome page if you don't know about it yet. You may get lost in here. lilienthalengineering.com/100-amplifiers-chapter-1/100-amplifiers-part-2-1945-54

    • @ElPasoTubeAmps
      @ElPasoTubeAmps  6 ปีที่แล้ว +1

      Absolutely awesome site. Thank you. I could spend a week looking at the schematics.

  • @MiddletonGuitars1
    @MiddletonGuitars1 9 ปีที่แล้ว

    I mean lift just the output leg if the cap