Other Amps

What happens to an amp that isn’t used for years?



Its a thing of Beauty.

I received the Brimistor Equivalents and I got one which matches the CZ4 closely. I held off to give you instructions as I first wanted to see if I actually have a replacement for the Brimistor in you AC50 from what I received.

I rewired my test amp output stage to resemble a vox AC30, as closely as possible, used EL34s and set B+ at your AC50 value using a Variac even removed negative feedback. It is as close as it can get to an AC50 and runs at about 50W too nicely.

I then added the Brimistor replacement and it works great. In my case I have less hum and I already had very little. I will work out later what your expected hum improvement will be etc before I send it to you.

I just attach the switch-on switch-off curve with the Brimistor in the High Tension B+ power line. As I already mentioned, it is a thing of beauty. The red is the supply side of the Brimistor, and the yellow line is the Brimistor connected to 40uF Capacitor side aka B+. As you can see there is a nice slow ramp-up to B+, and you can see the red displaying a solid 120Hz noise which creates the solid part as it oscillates. The Yellow B+ side is completely smooth and slow ramping and do not loose track as the wrong component would do and brown-out the amp at a too low voltage on B+ sometimes an hour later or immediately. You can see from the yellow line that there is now a long overshoot which is due to the capacitor internal resistance heating up due to inrush at switch-on. This is now WAY less than what it is if I just hard connect the capacitor to the transformer and bridge as you have in your AC50. It clearly protects the capacitor and clearly takes the place of a standby switch as Denney designed it as originally.

I just want to completely characterize the Brimistor equiv and carefully compare with the datasheets I have on the CZ4, but I tested it and played through the amp and it worked fantastic without the need for a standby.

I will post much more later before I send it to you as I need a few measurements from you to make sure my faux "AC50" is the same as your AC50 before I send it.


This is all very exciting, can’t wait to get that little bad boy in there.

Just let me know what you need me to measure and I’m on it.


The only drawback is that the one that is closest to the CZ4 is an ugly white color. Not going to look nice in your vox. I will have to go over the other lot I got and see if there isnt a black or dark redbrown version. It works spectacular, but doesnt look good. It must have been "happy day" in the factory back then when it was made. But these are custom made components, so I guess that is expected. Maybe it had to suit the ship's decor or something.


“Works spectacular” sounds good to me, I can always give it a lick o’ paint


This one looks better and tests the same. You can see my measurements on the paper underneath it. I cleaned it up a bit and it has the same white base underneath it seems, so I leave it like that. It looks better than the white one to me.


38mm each.

If you want to change their appearance as you indicated, remember they should be conformally coated which is a different process all together. I suggest you leave it as is, as they run hot by design and paint wont last and will start smelling etc. Ceramic or conformal coating is what must be used on these. Just use it as is. Also make sure you do not use tin solder with ultra low lead like the ROHS "Osha Cowboy" stuff. It will dry joint and grow tin whiskers on leads operating at elevated temperatures. Only use old-time high lead Rosin core.


Deke: I want to send you the Brimistor Equiv the following week. I just need a few measurements so I can verify what I test here is what happens in your amp.

Currently you have a 10 Ohm in place of the Brimistor. Please trace your circuit and make sure the resistor is in series with the bridge as on the schematic else the measurements will be futile.

If for any reason you dont want to do these tests and just want me to send the brimistor without double checking it is ok too. I am 90% sure it is the correct one.

You need to replace the 10 Ohm with a 47 Ohm and a 100 Ohm in turn and record the DC Voltage across the resistors. As an added feature measure the dB hum when the amp is fully on for each of the three following resistors after you taken the measurement. Make sure the background noise is the same and that the tablet is the same distance position from the amp. e.g press the microphone against the grill at an exact same point or anything repeatable.

As usual, be careful, all multimeter leads are at B+ during this measurement. Touch nothing when on or when not discharged.

Also bring it up slowly with the variac during the procedure of measuring.

  1. Measure the voltage across the 10 Ohm resistor and check if it is still as you measured before by connecting the multimeter across it and reading the DC Voltage. Wait till the amp is fully on. Let's say 1-2 minutes. Take dB reading. Switch off discharge.
  2. Solder a 47 Ohm in place of the 10 Ohm currently there, at the place where the Brimistor should be installed. Connect multimeter probes over 47 Ohm resistor. Switch on and record the DC voltage on the multimeter. You have to wait till the amp is fully powered on to record the reading. Take dB reading. Switch off discharge. NOTE: It should not happen, but if Voltage across 47 Ohm resistor at any time exceeds 8V after switch-on, switch off and abort measurement.)

  3. Solder a 100 Ohm in place of the 10 Ohm currently there, at the place where the Brimistor should be installed. Connect Multimeter probes over 100 Ohm resistor. Switch on and record the DC voltage on the Multimeter. You have to wait till the amp is fully powered on to record the reading. Take dB reading. Switch off discharge. NOTE: It should not happen, but if Voltage across 100 Ohm resistor at any time after switch on exceeds 18V, switch off and abort measurement)

  4. Replace 10 Ohm resistor as it was originally.

This will give me the required I/R curve from which I can relate all my measurements here accurately as the curve will characterize your power section steady state conditions.

This is the last time you will have to take diagnostic readings. the good is that this will give you a Brimistor and restore your Vox. Not much to ask for that. After the Brimistor and the other missing capacitors you identified are installed, you can go all the way and install BY100's as they will be fully protected against transients.


Ok cool, I’ll take care of this today. I’ll replace C9 and C23 then get onto this. I should have the results by mid afternoon UTC.

The solder I have was given to me by a TV repairman who came to fix our TV in the mid-70’s when I was about 13 or 14. Is that Olde-Tyme enough?

New diodes are not here yet so they will have to wait until new week.


10Ω 10w Resistor

1.4vDC across thermistor points, image shows hum measured from approx 3" using DeciBel X Pro app, recorded over approx 10sec.


47Ω 10w Resistor

5.8vDC across thermistor points, image shows hum measured from approx 3" using DeciBel X Pro app, recorded over approx 10sec.


100Ω 10w Resistor

10.9vDC across thermistor points, image shows hum measured from approx 3" using DeciBel X Pro app, recorded over approx 10sec.


Fantastic thank you very much! This is extremely helpful and very well done and I can generate the curve now.

At 100Ohm the Hum is 1:1 in ratio to your Hiss !! and only a little more at the lower values. That is already very low for a Vox. Voxes tend to hum a bit by design, but that sounds normal to me. No sweat we can get it still lower if you wish.

Last thing. Can you download the following APP. "Audio frequency Counter" by Keuwelsoft and more importantly. "Spectroid" by Carl Reinke: the latter is very important, and I hope it is available for IOS. I use Android.

With your 10 Ohm resistor installed. Start the amp and let it warm up a bit. Then Startup Spectroid and hold it at a similar position you did before. The app is integrating and it is necessary there are no other noises. All amp controls on ZERO.

The red curve is the peakk-line which you can ignore. With only the hum audible use "Spectroid" for about 10-20s ..no other noises else you have to restart. Take a screenshot and post it. Make sure you dont make noises taking the screenshot. Figure it out beforehand to do it softly.

What is expected you should find.

If your amp power supply section works correctly you should only get a single peak at 100Hz and none at 50Hz. If you get a 50 Hz peak also we have to trace that issue. It will be either heater circuit or a failing diode or failing bias regulation (which you said you fixed due to the missing capacitor you found). If you get strong higher order harmonics it might point to noise or harmonics on your AC supply. I already checked and your supply does filter 100Hz directly proportional to the resistance values which is what is theoretically expected. If you find harmonic peaks >100Hz that do not filter or 50Hz peak present something else is going on.


I ran the recording I did yesterday through a spectrum analyser within Logic Pro, will this tell you what you need to know? It looks like the peak is around 150Hz.


Deke, yes that will help a lot. What you show in your previous post doesnt show these curious shifts in the app spectrum analyzers I see. Their windowing is seemingly not centered and the displayed frequencies are shifted higher than actual. Pretty gross oversight.

What you show with HEQ is exactly right. Can you please make a video or just three photos of the HEQ interface as you move from 10/47/100 by playing your previous sound sample through that if you are already set up top do it. If it is too much hassle, I will run it through Mixbus here and use one of the FFTs. I just ask you to do it since what you get in your previous screenshot is exactly right and has integrity, so I want to continue with what you have. I now just want to see it for all 3-values.


Pending the three photos or a video; What you see there are the even and odd harmonics of full-wave rectification. A single sinusoid has only a fundamental frequency, but since the diodes only charge the capacitor at the peaks of the full-wave rectified (100HZ) signal, the current is chopped up and only conducts at these peaks and the charging current is not sinusoidal. therefore since it is not sinusoidal, it will create even and odd-harmonics on both sides of the bridge. That is why I want to see the spectrum at the right frequencies, that seemingly the apps miscalculates. The Android and such apps shows it at wrong frequencies, but your HEQ shows it 100% right. Generally these harmonics will be filtered by the supply capacitor. I need to see what the 100Hz peak does w.r.t the other peaks of odd and even harmonics of the original 50Hz signal. It seems on first glance the harmonics pass through, which should not happen, but that the fundamental (of the rectifier) namely 100Hz is filtered. So it would be nice to see 3 screenshots of HEQ for each of 10/47/100.

Regardless of this, your hum is clearly 1:1 with your hiss level which is actually very low for a vox in general and acceptable. But we can get it better if you so choose, although it will take a bit of time.

QUESTION1: Were the soundsample taken with the missing bias capacitor C23replaced and installed or not ? I am not sure at which stage you took the sound sample.

QUESTION2: You also found tracing your circuit that C26 is missing. C26 should help filter the harmonics you see in the spectrum created by the rectifier bridge switching current so that these harmonics do not pass on to the Bias circuit. If you replace C26 it must be a class Y capacitor and not just any old Sprague or definitely NOT a class X capacitor as you need to install a capacitor there that wont short as a fault condition. If C26 ever shorts, it blows your Mains transformer. Therefore it needs to be Class Y as it will fail open. See if you can get a 1500V-2000V Class Y capacitor at value equal or slightly bigger than the schematic. If you put e.g. the usual old Sprague that everyone banks on in there it will eventually eat your transformer. Type Y only.

I have the suspicion that Denney added that capacitor for exactly the reason you see the harmonics appear in the output transformer, likely leaking through the bias circuit to the output tubes. The harmonics of the hum should not be there.

As I said before, I clearly see no 50Hz peak in the spectrum, so clearly you dont have heater induced hum as the 50Hz from the heaters do not appear in your output transformer.

Anyway, I just wanted to justify why I want the spectra you measured for the sound sample at the three values 10/47/100


Having trouble posting images for some reason so I sent privately.


QUESTION1: Were the sound samples taken with the missing bias capacitor C23 replaced and installed or not?

Yes, after I'd replaced C9 and C23.


Having trouble posting images for some reason so I sent privately.

– Deke Martin

On GP, I always have to add an image before I write text, else the image doesnt load. And I cannot add more than 1 image per post. It is understandable, I guess as a lot of images can very quickly eat space on a server so I guess it is a way to make us post only images that are absolutely necessary and to rather use links for multiple images.

Thanks, I received the 3 photos you sent privately.

Register Sign in to join the conversation