s-petersen
Scott
The step network seems excessive, and there is some ringing, I have to look at the schematic and get back to you, at least both channels are similar. Normally you should not do full power square waves, it's pretty hard on the amp.
Thank you! Yeah my first real restoration project here, so I did not know that. I will make a note to not run square waves at full volume. Luckily that was just for a few seconds.
Not sure if this makes a difference with ringing specifically, but this was just a quick test within 5 minutes of the amp being turned on. I did not let the thing run for half an hour.
What does "the step network" mean?
s-petersen
Scott
I looked at the schematic, and there is no step network in your schematic, usually the step network is in the first tubes plate circuit, a resistor and cap in series across the plate resistor, to reduce the first overshoot in the square wave and it also does some rounding of the square wave leading edge.
You may want to change the 180 ohm resistor to 1K and check the 10k low powered square wave again.
The ratio of the 820 pf cap and the 180 ohm resistor, is very different from the Mark 3 the amp is copied from.
You may want to change the 180 ohm resistor to 1K and check the 10k low powered square wave again.
The ratio of the 820 pf cap and the 180 ohm resistor, is very different from the Mark 3 the amp is copied from.
Last edited:
Here's my setup:
top scope is the probe on output of signal gen, frequency counter is clipped there too;
outputs from the amp via regular speaker wire to 100W load resistors mounted on the heat sink;
lower scope leads across the load resistors: top is left ch., bottom is right ch.;
100Hz:
1kHz:
10kHz:
20kHz:
Under these conditions I can see oscillations at 10k and 20k. Next step: replace that 180Ohm resistor for a suitable 1K resistor, just like everyone told me a few times already.
Can anyone approximate by eyeballing what is the wattage of that resistor? My guess is somewhere between 1 and 5 watt?
s-petersen
Scott
I believe on the Dynaco schematic it was 1 watt
yep, thats fine. Just wasn't sure what the level was running at since an earlier post said something about "full". I wasn't sure if that meant full power or what. Running it way up will skew what you actually see on the scope.
The resistor off the plate won't actually be handling all that much power, but its dealing with a lot of voltage. The higher watt resistor was probably picked more for that reason than any other.
The resistor off the plate won't actually be handling all that much power, but its dealing with a lot of voltage. The higher watt resistor was probably picked more for that reason than any other.
Yes, you understood it correctly. The previous set up was to look at the output at full amplifier output and at some random point at less than full (say 75%) while the load was a set of speakers;
My last post with pictures was at the amplifier output of 2 volts p-p as you suggested. I absolutely agree that my previous two setups did not show as clearly what was going on. (I'm assuming you can click on the thumbnail and it brings up enlarged image) Your suggested setup shows ringing clearly at both 10k and 20kHz, so thanks for the suggestions!
Sorry, I'm confused by the "resistor off the plate" comment, so I wanted to measure twice before I splice, so to speak. I am replacing the 180 ohm resistor (per channel, with a 1kohm) that goes from the 16 Ohm tap output to the cathode of the pentode on the 6AN8. Is that correct? I'm not confident in my ability to eyeball the wattage of the resistor, but Scott suggests Dynaco schematic has it as 1 watt. Lengthwise it measures around 17mm. Here: http://music.codydeschenes.com/wp-content/uploads/2013/06/abres.png it says 15mm is 1watt, 18 mm is 2 watt. I do have 1k 1watt metal oxide resistors on hand that I could use in that position, does that sound about right?
Last edited:
Oh, I found a very useful guide on Allen Bradley resistors that also shows dimensions and voltage ratings, if others find it useful: https://www.hificollective.co.uk/sites/default/files/allen_bradley_data_sheet.pdf According to this (pg 2 ) 17.48+/- 0.79 is a dimension for a 2 watt resistor.
s-petersen
Scott
Using a 1 watt or 2 watt resistor is OK either way, The Dynaco schematic was 1 watt, but 2 watt will not hurt anything.
swapping with the 180 ohm is the correct thing to do.
swapping with the 180 ohm is the correct thing to do.
I was thinking of something else that isn't actually there. For some reason I thought the Dynaco had a resistor and cap from output tube plate back to the pentode and thats what you were talking about. Disregard. Its got just a 390pf cap from the screen.
Ah thank you for clarifying. Now if the 1k resistor does not work well enough for some reason (I'll try that first), does putting there a 1k pot temporarily just on one channel to fine tune a resistance that will give the flattest response an option?
The feedback resistor value is determined by the amount of feedback it provides, in this case 20dB. You can check the amount of feedback with the procedure explained in this thread, comment #48.
https://www.audiokarma.org/forums/index.php?threads/citation-ii-hack.471242/page-3
In this case 1K should provide the correct amount.
Only then should the square wave response be fine tuned by adjusting the value of the phase-lead cap across the feedback resistor. Start with the original value in the schematic.
You generally don't tinker with the feedback level as a way of adjusting the amp's frequency response.
180Ohm replaced with 1KOhm. I put new waveform followed by the old waveform (with 180Ohm resistor) for ease of comparison. Here are the results:
100Hz:
- 1k 
- 180Ohm res1KHz:
- 1k 
- 180Ohm res10KHz:
-1k 
- 180Ohm res20KHz:
- 1k 
- 180Ohm res
Last edited:
It certainly looks more stable, much less ringing. The 1K resistor is taken from the 16 ohm tap, is that right? Did you leave the orginal phase-lead cap across it? More importantly, did you measure the amount of feedback? It's really important to ascertain that there is, in fact, 20dB of feedback before you do any more tests or fiddling with the components.
s-petersen
Scott
It looks much better, If the gain is reasonable I would leave it that way
Last edited: