Isolation transformer installed....hum

[pontiacpete]

I just finished wiring in an isolation transformer. The original schematic is below and a 2nd one with the changes to the circuit. As you can see this amp had a floating ground. I changed all the grounds to the chassis. There is a slight hum with the volume down, but certainly not loud enough to be a concern.
The hum that it has is affected by the volume, is there something that could be change to make it quieter?

[Dave H]

I don't see the floating ground connected to the ground. It was connected by a parallel R and C before. I'd just connect it to ground (chassis ground)

[pontiacpete]

Which is what I did. I know my drawing is really lamesorry
But all ground points, 'chassis' or 'floating' are now attached to the chassis.

[J M Fahey]

Why do you use different ground symbols in your corrected schematic?
That is at least misleading.

What happened to the 100k//.05uF ground separation network?

In principle, when you add an isolation transformer to a widow maker (good), you just connect what formerly went straight to the mains plug, to the floating isolated mains voltge secondary winding, without any othr change, and trynthe amp.

Only after that you may start trying to improve it.

In a nutshell: one end of floating secondary winding should go to the rectifier anode as before, the other end to the former floating ground, which is the main one.

The preamp pentode used to have a 150k cathode resistor which you removed.

Please redraw the schematic using 2 different symbols and where proper if you keep the separation network or using the same symbol everywhere if you join them.

Thanks.

[pontiacpete]

That was my first move. I put in a fuse and moved the switch to the primary of the iso. But since the other side the switch was connected to the floating ground as indicated by the symbol on the original schematic. ONe end of the heater string was connected to the floating ground also. Anyway when I disconnected the switch I wasn't getting and heater voltage. That's when I tried removing the 100k/.05uf and brought everything to chassis ground.

I didn't remove the 150K cathode resistor. That just says the number 7 for pin 7 of that tube. Is that what you are talking about?

Thanks for the help guys!

[Enzo]

Adding the iso improves the safety, but not the circuit. It hums because it was never designed to be very quiet. Your B+ is half-wave rectified first off, but also it is single ended, so ther is no ripple cancellation in the power stage. The B+ for the power stage comes right off the rectifier and first filter. You want less hum, add a new first filter and a resistor between the rectifier and that tap to the OT.

[pontiacpete]

Thanks Enzo that helped alot to reduce the hum. I added in another 68uf cap and a 1k resistor.

I thought we might be talking about the position of the volume switch or something else

[J M Fahey]

the 100k/.05uf is also a cathode resistor.
Just follow the path from (real/main) ground at the main filter cap negative leg and first pentode cathode

It does not affect gain because they cheat and apply signal from grid (jack hot) to cathode (jack "ground") so pentode runs at full gain, and cathode resistor is so large (150k) that even if you plug the mains plug backwards and supply negative now is "hot" (oh yes it will), and main ground now is hot, you won´t die when touching the guitar strings because current is limited to tingly but not deadly 1mA (roughly 150V DC across 150k resistor).

That´s why not many people died when using these things, although it was still deadly if you touched an internal hot point, even with just one hand and the other in your pocket, because you are still grounded through your feet.

So they lowered death rate from 50% (you have 50% chance of plugging it backwards) to 1 in 100 thousand or 1 in a Million ... which is far worse than it sounds because these things are portable and everytime you unplugged and replugged you pulled a new lottery ticket from the basket .... eventually somebody would win the prize.

I distinctly remember being a kid and *a lot* of stuff "tickled", including room fans, desk lamps and other stuff.

Of course they were damaged but since it was "a reality of life" nobody cared too much.

Even today , the river near my house floods, no big deal any more where I live because they built a huge wall along the river with automatic floodgates but on the other side, a poor people area, it still does and everytime we read in the papers about a couple people died trying to pull from their house the TV set, the fridge, etc. ; their most prized possessions.

Of course modern electronic stuff has hot chassis SMPS inside, appliances run from straight 220V, so they are safe in normal circunstances but when soaking water by all holes they become deadly.

The point is that even today, with all the electrical awareness available, still (some) people dies.

FWIW now when a flood is coming (meteorologists announce them with precision, 24 or 48 hour earlier) , Electrical Company plain cuts power in the endangered areas, period ... against the loud complaints of inhabitants (who should know better,then you see them crying on TV) .

This is my very street corner (I live on the left street) just 20 years ago, before the wall.


in fact I used to swim there, olympic style, under the applause and encouraging chants from neighbours, as well as sail inside some old discarded cabinet.

[pontiacpete]

Interesting stuff about you neighborhood. Like that picture!

Do you think if I put the 100k /.05 cap back in it will work any better? Still has a bit of noise . I'm wondering.
Or is this just a safety issue?
Would adding a cathode resistor on the first cap tame some of the noise?

[J M Fahey]

No, don´t think it would change much.
When I first noticed that I thought that it would affect bias, but then after further analysis 2 or 3 things came to mind:
1) they "cheat" because signal is applied between grid and cathode, and like Enzo often says, "that´s all a tube cares about" so then:
2) I considered it (from the DC point of view) like something similar to LTP PI "long tail resistor": it does not affect bias because tubes are self biased from their own shared 470 ohms resistor and grid reference 1 M resistors are also referred to that same 470 hms resistor; all the long tail one does (besides providing a relatively constant curretbsource) is waste voltage.
As in: if PI is fed, say, 350V DC but loses, say, 60V across tail resistors, it only has 350-60=290V ... no big deal.

Now on a transformerless amplifier, where best case you have 150/160 V available, then that loss is more significative, I remember common tube pre effective supply voltage is only 60 to 90V ... explained by loss across the 150k separation resistor.

To minimize such drop they use a very high 1M looad resistor.

So I guess now you have a higher supply voltage to feed the preamp.

I guess you can spend that bonus into adding extra RC filtering to the preamp.

Besides that, not much else to do; as Enzo said you have half wave filtering andn single ended supply, it comes with the territory.
You´ll survive

IF you wanted a hum free amp, you would have built an AX84 or something

[g1]

Is there any reason it needs to be a half wave rectifier?

[Dave H]

Is there any reason it needs to be a half wave rectifier?

If you use a bridge one end of the series heater string won't be grounded. Will that cause hum?

[pontiacpete]

I tried to put bridge rectifier in place of single diode. Attached the secondaries of the iso and the + to the B+ and - ground along with one end of the heaters and still hum.
Again this hum is directly affected by the amount of volume. Wouldn' t a heater hum exist with the volume down and not get louder with the increase in volume?
Perhaps I need to put the bridge before the half wave rectifier. Would that help?

I've tried to do a proper drawing but so far I'm inept at computer programs that do this. Such a shame. Spend several hours and still the 'white out' solution if I had some, would be a lot more easier.

[Enzo]

It depends on which tube the heater is injecting the hum into. heater hum is not generic, it will affect individual tubes in individual circuit nodes. Typically it will most affect the high gain input stage.

With a bridge, what would be the point of having the diode?

[nickb]

If you want to add a bridge, here is how to do it. This gives you a couple of benefits. It makes the current in the iso transformer symmetrical so it is less noisy, runs cooler and avoids and danger of saturation due to the half wave rectifier. It also supplies the amp with 120Hz DC so you get less ripple. You MUST keep the original diode in place BTW.

I suggest that you move the 300 ohm resistor to be first in the high voltage side of the chain. This will reduce the hum voltage on the heaters and with any luck will reduce your overall hum level. If hum get's into the very first stage (12AU6) and since that is before the volume control that part of the total hum will change with the volume level.