Little Heathkit SA-2 hum trouble - FIXED -

[Brice]

Hi guys,

I bought a couple of years ago a Heathkit SA-2 amplifier, but I don' t know why.
This is a full integrated stereo amplifier with phono sections. A nice little amp to have in an office.

I then decided to dive into it and restore as it is extremely clean and beautiful, with the same iron than the UA-1 so I knew it will sound nice.

I received it untouched, and decided to try it. The amplifier did work nicely, just with some audible hum.
It has no shield near the input jacks, a big huge caps was added on top of the chassis which didn't look original.

Other than that it looks super nice.






I started by testing the resistors and caps and they were all out of tolerance, so by the end of my investigation, I realized most of the components were bad so no point to keep some old parts (I've been burned before), and change all.

So here is the results after a rebuilt:




But I had hum. A bad one. Noise was everywhere but mostly at the 6AN8 stage.
I swapped all tubes around, did nothing.

I started to do the usual tricks to find the root cause.
I've started with the PS as it was a weird 60/120 modulated signal.

I then discovered that the PS filtration was very poor as is, Even with the Authenticap cap I installed with an increase of capacitance, with a V+ ripple of several volts to the OT CTs. Not good.

When I debug an amp I always use a SS rectifier so I don't wait. And ramp slowly with my variac.

So I added more capacity on the B+ and B+2. It did help. But not enough.

I ended by adding 2 identical decoupling stages before the regular power supply, of 50R 56uF.
That did the trick nicely with a ripple now of 100mV.
And I did add a 120uF at the B+2 (after the 4.7K), which also greatly help.
All this lower the B+ a bit, but with the SS plug, it wasn't a problem.

But the hum was still there, around 7 to 14 mV at the speakers. So I ruled out the PS.
The ground was nice to begin with, I changed just a couple of path, like the cathodes return directly to the PS can, the feedback reference in the OT to the input reference, that sort of things.

No change as I was expected, unfortunately.

So something was picking up noise. I decided to use shielded wires for each EL84 grid. These wires were long so a potential problem. It did nothing.

I then shielded the 6AU6. No change. Then the 6AN8, nothing. The 12AX7 were already shielded.

Now I was getting mad at that thing. It should make no noise as it was and I spent a lot of time already on it.

At the end, I just changed the 2 identical NU 6AN7 with some others I had and... the hum disappeared.
The amp was totally silent.

I was mad at myself not trying other tubes but to my defense, I did swap them with each other but they were both equally bad. I did not expect that.

So all this hum investigation (Thanks Gadget for your advices) lead me to some mods which in fact did improve the amplifier, so I wanted to share them. Here they are:

1. Separated the common 4 tubes cathode resistor 100R / 50uF into 2 x 250R / 50uF to each channel EL84 cathode resistors. Channels are now independant.
2. Added UF4007 diodes upstream the rectifier.
3. I use a SS rectifier and not a GZ34 any longer: save heat and power for the PT.
4. Added 2 stages of 50R 50uF before the regular supply. Ripple much much reduced from 2 volts to 0.1V
5. Added 120uF to C24 so now 160uF instead of 40uF.
6. Added a thermistor.
7. Added an internal B+ fuse.
8. Changed C16/C41 from 0.1uF to 0.22uF.
9. Shielded all long grid wires to each EL34.
10. Added shields to all driver tubes.
11. Added tube retainers to each EL84.
12. Added a snubber cap across the power switch.

I am listening as I am writing this, and this little amp is wonderful. It's a nice little guy.
Some last pics.

Cheers,
Brice.

 

[radioactive]

i had a pairvof ua-1's and knew these used the same opt's .The ea2 monos also used the same opts .So if anyone wants to clone a ua1 or 2 without the price of paying for those find youself 2 ea 2 amplifiers and your in business.

 

[fxer1982]

Real good description of work done to get where your very happy with it...nice looking amp, must sound great.

 

[spantou01]

i had a pairvof ua-1's and knew these used the same opt's .The ea2 monos also used the same opts .So if anyone wants to clone a ua1 or 2 without the price of paying for those find youself 2 ea 2 amplifiers and your in business.

I bought an EA-2 for this purpose and it was fitted with a 51-30 output transformer with a smaller stack than the desirable 51-29, so be sure what output transformer you get.

 

[zackthedog]

That's a beauty! Very nice restoration and a cautionary tale as well! ;-)

 

[spantou01]

• Added a thermistor.

• Added an internal B+ fuse

Nice description. I recently acquired a Heathkit SA-1, so this will be helpful.

Could you provided more details about the:
thermistor - what and where installed?
B+ fuse - where in the circuit and what type and amperage?

 

[Brice]

Thanks guys.

I used a CL80, 3A, 47R thermistor, installed next to the rectifier.



I install the fuse on the B+ going to the CT of each OT. I used a 500mA fast fuse.

 

[gadget73]

Some of those pentode/triode tubes are just funky. So far as I understand the only one specifically intended for audio use was the 7199, so it may just be a case of there wasn't a ton of attention paid to hum because in their main application it just doesn't matter. Luckily they tend to be pretty cheap so its not that expensive to buy several sets in case you have to swap them to get quiet tubes. I think all of my 6AN8's were from the $1-2 bins at Kutztown.

 

[nerdorama]

i had a pairvof ua-1's and knew these used the same opt's .The ea2 monos also used the same opts .So if anyone wants to clone a ua1 or 2 without the price of paying for those find youself 2 ea 2 amplifiers and your in business.

Exactly what I did. They were rat rods so I gutted them and only rebuilt the power amp sections. Put on a blank powder coated face place with a power pushbutton for a nice simple look. Also added a choke in the power supply since I had plenty of space. Mine have Stancor 51-30 output transformers. They stabilized fine and sound great.

https://audiokarma.org/forums/index.php?threads/choosing-a-good-el84-amp.1022595/#post-16211700
https://audiokarma.org/forums/index...t-ea2-3-amps-with-stancor-51-30-opts.1010343/

 

[gadget73]

kind of weird how many good power amps are mated to not-great preamp sections. I think thats part of my bias against tone controls. They aren't all bad, but some of them sure do suck.

 

[spantou01]

Some of those pentode/triode tubes are just funky. So far as I understand the only one specifically intended for audio use was the 7199, so it may just be a case of there wasn't a ton of attention paid to hum because in their main application it just doesn't matter. Luckily they tend to be pretty cheap so its not that expensive to buy several sets in case you have to swap them to get quiet tubes. I think all of my 6AN8's were from the $1-2 bins at Kutztown.

Check out the 7687 for an “improved 7199-type”. Here is the Sylvania datasheet: “Type 7687 features exceptionally low hum and noise output. Although basings differ, Type 7687 is similar to Type 7199. Type 7687 is an improved low noise version of Type 7199”. It was only made by Sylvania, so they may be a little biased

 

[dcgillespie]

Brice -- Nice job! If I may however, let me offer two points for your consideration:

1. As I discussed in my thread on the SCA-35, all of the tri/pent tubes are rather notorious for either producing or developing an unacceptable level of hum from (most often) the triode section of the tube at some point throughout their usable life, when the heater is operated from an AC source. This problem can be greatly reduced by using a series dropping resistor to drop the heater voltage to a level that is 10% below the nominal rated voltage for the heater. You might try this with your hummy tubes -- it will likely help with such tubes.

2. Splitting a common cathode bias resistor for 4 output tubes in a push-pull stereo amplifier is a popular upgrade, and its does help to ease the balancing requirements of the tubes by dividing the matching needs into two separate pairs. But this benefit only extends to the tubes under quiescent conditions. Under dynamic conditions, this move actually hurts the performance of the two amplifiers because it prevents them from acting together to effectively regulate the bias for both pairs of tubes. This happens because the stereo signal is rarely identical at any given moment in the two channels, so when connected together, the unequal current draw between the two channels -- over time -- tends help keep the bias steady, which then helps to maximize the transient power response of both channels. In deed, just as with the Dynaco ST35, if each channel is measured separately using a sinusoidal test signal, they will each individually produce (about) 17.5 watts of power -- because in his case, the undriven channel is acting to regulate the bias of the driven channel. With both driven however, power output drops to just 12 watts per channel because in that scenario, the signal is identical in both channels, so no bias regulation effect can occur -- even though the cathode pairs of both channels are connected together. This drop in power is slightly due a drop in B+ when both channels are driven, but is most largely due to the large increase in bias produced in his scenario. To the point then, when the connection between the cathode pairs for each channel is severed, then no bias regulation can ever occur, and the channels are condemned to a maximum power output of just 12 watts under any condition of operation. It was this very effect that led to the development of the EFB biasing concept that allows the small Dynaco amplifiers -- and others with the same type bias design -- to actually develop full power output in both channels at the same time, with otherwise identical signals applied to both channels simultaneously. This is not a pitch to install EFB in your SA-2, but merely an effort -- by way of describing its reason for development -- to help explain why the popular modification of this discussion, isn't always as good as its always advertised to be. The effect is not as bad in the Heath amplifiers due to the higher reflected impedance of their OPTs -- but the effect is still alive and well even in their designs. Bottom line then, what is often described as a cheap economical biasing system, actually has some method to its apparent madness........

Dave

 

[gadget73]

Check out the 7687 for an “improved 7199-type”. Here is the Sylvania datasheet: “Type 7687 features exceptionally low hum and noise output. Although basings differ, Type 7687 is similar to Type 7199. Type 7687 is an improved low noise version of Type 7199”. It was only made by Sylvania, so they may be a little biased

quick poke around implies these aren't' very available either though. What I want is a fancy audio grade tube for a dollar that I can get everywhere. Then again there are expressions about people in a particular place wanting ice water

 

[Brice]

Thanks Dave.
I didn't know about hum and heater voltage correlation.
And I understand the second point. thanks for your detailed explanation, but I like to keep the amps separated as much as possible.