Isolation transformer installed....hum

[pontiacpete]

Here's some pictures.

I'll try it around the apartment to see if there is any difference.

[olddawg]

Honestly... I have never seen a series filament amp that didn't have substantial hum. I always assumed it was because you cannot have a self canceling twisted pair for your heater wiring at 6vac. You have a filament string that's 120vac at the ends radiating into everything! Probably 20 times the hum just by design.

[Enzo]

It hums for a million reasons. First series string sets are all low end basic units: table radios and cheap record players. The signal path is near ground in terms of signal, but the heaters can be 100v above ground, so I suspect more H-K leakage than lack of twisty. For that matter, there is no reason you can't twist the heater wires. The wires don;t know if you are series or parallel. Just like a hummy Champ, these inevitably take the B+ for their single ended output stage off the first filter, so that leaves some hum. And for that matter, it is half wave rectified B+. One thing we have corrected here.

[olddawg]

It hums for a million reasons. First series string sets are all low end basic units: table radios and cheap record players. The signal path is near ground in terms of signal, but the heaters can be 100v above ground, so I suspect more H-K leakage than lack of twisty. For that matter, there is no reason you can't twist the heater wires. The wires don;t know if you are series or parallel. Just like a hummy Champ, these inevitably take the B+ for their single ended output stage off the first filter, so that leaves some hum. And for that matter, it is half wave rectified B+. One thing we have corrected here.

Yep and single ended amps are hummy by nature as well. That's why so many of the modern 5 watters have DC filaments.

[Enzo]

It is important to keep in mind that hum is not generic, and can come from many causes. And EACH cause has its own cure. Adding filter caps does nothing to improve grounding hum, and changing the grounds won't reduce ripple in a weak cap. Hum radiated from AC heater wires is not the same as hum from H-K interaction. And so on.

[nickb]

It's hard to see but I don't think there is a problem with the way the bridge /filter caps are wired as far as the grounds. I prefer to keep that part of the wiring tight so as to minimize the area of the loop transformer-> rectifier-> first filter -> rectifier-> transformer to avoid radiating noise. Twisting these wires together helps to reduce the field around them in so far as that is possible.

There are two problems with the heaters. As Enzo & olddawg have said these are not twisted and carry a big AC signal. Also, current spikes from the rectification process will get onto the heater wires and so propagate throughout the amp. Now, since we have the heaters after the bridge and a resistor, we can easily get a DC heater supply and avoid all that misery. Add a 100uF cap of at least 200V across the heater chain i.e. +ve to the 50C1 and -ve to the bridge rectifier. This will give you very close to 70VDC at about 1.5Vrms ripple without changing the 300 ohm dropper. Also it will shunt quite a lot of the rectifier noise to ground and help to keep it off the heater wiring. Your 100uF needs to have a ripple current rating of at least 150mA.

It looks like the yellow wire in the top left is the input signal from the jacks and it goes to that tag strip. That might be vulnerable to noise. It also looks like the adjacent tag is ground. As an experiment, simply short the signal to ground at the tag strip as see how much it reduces the buzz. If it does try keeping the wire close to the chassis or screening it. Also I wonder about the ground path to the jack - how is that done?

Note I use buzz to refer to the spikey high frequency component of the noise and hum to the low frequency stuff i.e. primarily 120Hz or 60Hz.

[pontiacpete]

Wow!! that worked like a charm. Thanks Nickb!

First, I went ahead and installed 2 tip shorting 1/4" input jacks with a shielded wire and 68k resistors and a 1M a la fender style. I always knew I would have to address the noise problem when an instrument isn't plugged into the input. It was ridiculously loud. But still the hum with volume increase

Then I put the 100uf / 400v cap across the rectifier, + to + on the rectifier and (-) to (-) or ground, and the amp is dead quiet. Amazing.
I was almost ready to give up. Hope my cap is good for the job.

Thanks again for the help!!

[nickb]

wow!! That worked like a charm. Thanks nickb!

First, i went ahead and installed 2 tip shorting 1/4" input jacks with a shielded wire and 68k resistors and a 1m a la fender style. I always knew i would have to address the noise problem when an instrument isn't plugged into the input. It was ridiculously loud. But still the hum with volume increase

then i put the 100uf / 400v cap across the rectifier, + to + on the rectifier and (-) to (-) or ground, and the amp is dead quiet. Amazing.
I was almost ready to give up. Hope my cap is good for the job.

Thanks again for the help!!

The cap needs to go across the heater chain not the rectifier- you'll burn out the heaters - fix it now.

Add a 100uF cap of at least 200V across the heater chain i.e. +ve to the 50C1 and -ve to the bridge rectifier.

[pontiacpete]

Too bad the heaters can't handle that, it was really quiet across the rectifier.

So, hope I have this right now. I inserted the cap between the rectifier + and the 300R resistor. (+) end to the 300R and (-) to the bridge.
This cut the heater hum in half, which is more than acceptable for this amp, if this is correct, of course.

[nickb]

Good
Well, that tells you that ripple on the B+ supply is an issue. Increase the first filter cap (after the single diode) to 100uF.

Anyway more important - I goofed - sorry. I forget to check the dissipation in the 300 ohm resistor - it's close to 10W. This is too much for a 10 watt resistor. You'll need a 20W one (or combination to the same effect) and locate it somewhere where all that heat will not be a problem. The power goes up because of the ripple current in the capacitor.

[pontiacpete]

The 100uf cap is getting a bit hot..
What kind of combination can I use? 2 paralleled 680R 10W resistors?
I definitely don't have a 20W 300R resistor.

[nickb]

The 100uf cap is getting a bit hot..
What kind of combination can use? 2 paralleled 680R 10W resistors?
I definitely don't have a 20W 300R resistor.

Two 680 10W in parallel should do. Check the voltage across the heater chain, 74V is the aim.

If the cap is getting hot, then it's likely it's ripple current rating is too low. 150mA is the minimum at 120Hz. For a decent cap life want to go for a rating of 300mA or more.

[pontiacpete]

Looks like I will have to order the cap. I pick this up when they come in.
Thanks

[pontiacpete]

Got the caps and resistor in today and now running them in the amp.
Hum is reduced to a reasonable level for this amp, but I'm a little concerned about the cap in series with the 300R resistor(50 watts), runs pretty hot.
I can't touch it for more than 5 seconds. This cap is rated for 430mA. Here's the datasheet:

http://www.mouser.com/ds/2/293/e-upt-876272.pdf

will this hold up?

[pontiacpete]

I guess I just found the answer to my question. Smoke started coming out of the cap.
Must not be good enough for this application.