a.c. heaters

[Brian Roth]

IIRC correctly, the PCB traces for the 6.3 VAC lines (heaters and panel lamps) have no connections to any of the various ground traces.

I was also puzzled, since the "original" EMI schemos showed AC on the filaments.  I also swapped a few different tubes in, just in case there was an errant Fire Bottle.


Bri

 

[abbey road d enfer]

IIRC correctly, the PCB traces for the 6.3 VAC lines (heaters and panel lamps) have no connections to any of the various ground traces.

Maybe they've left the heaters floating, which would explain the noise; in that case grounding one leg or a virtual center-tap would have fixed it.

 

[MicDaddy]

> shocks, is this stray capacitance through the transformer or is this having to do with a hot chassis

Amp with one 6V6 is "surely" not literal hot-chassis.(*)

Your shocks may be the "small" stray capacitance in the PT.

Or it could be failed insulation in the PT.

WHY haven't you installed a proper modern 3-pin plug already??

You know you should.

It would divert stray-capacitance current.

If you have a hard-fault in the insulation (some people hardly notice straight 110), the 3rd-pin would blow a fuse and PROVE you have a problem. (Minor, since any Champ replacement PT will do the job.)

Hot-chassis guitar amps have tubes like 50L6, 50A6, 25L6, 35L6--- the two biggest tubes will be 25V-70V heater, the first two digits. While such a line-up "could" still be transformer isolated, the whole reason for such oddball voltages is so they (and you) can be strung-up directly across the 110V wall outlet.

CJ's toy is odd in that ONE of the tubes "is" transformer isolated heater. This is not even a half-assed attempt at Safety (maybe third-assed). I believe this type was made UnApproved as soon as UL could re-write their standards to keep-up with over-clever designers and marketers.

I see, I had assumed with the B+ center tap and one leg of the ac heaters tied to the chassis that this constitutes a hot chassis.  But that it's entirely something else is just my misuse of terms.

I think the B+ center tap needs to be tied to chassis to complete the rectifier circuit? 

Should I leave the one heater leg to the chassis or float both? 

What's confusing me is that the B+ is some 500V+ with the center tap tied to the metal chassis as well as one leg of the ac heater, with installing a chassis to the 3rd pin earth has me kind of  . 

I think I want the B+ center tap to be tied right at the first filter cap negative leg?

The IEC chassis ground on it's own lug to chassis near the IEC inlet. 

Still not sure what to do about ac heaters.

 

[abbey road d enfer]

I see, I had assumed with the B+ center tap and one leg of the ac heaters tied to the chassis that this constitutes a hot chassis.  But that it's entirely something else is just my misuse of terms.

I think the B+ center tap needs to be tied to chassis to complete the rectifier circuit? 

It needs to be connected to "ground"  (Zero-volt). Then you may choose to connect  "ground" to chassis; that would be safe.

Should I leave the one heater leg to the chassis or float both? 

Certainly not leaving it float. You may choose to ground one leg or the other, or create a virtual center-tap with two low-value resistors (about 100r) or even a potentiometer, which in some cases may provide some noise adjustment.

What's confusing me is that the B+ is some 500V+ with the center tap tied to the metal chassis as well as one leg of the ac heater,

Which just means that HT zero-volt goes to chassis -chassis being ground) and the heater winding has one leg grounded. Nothing really unusual.

with installing a chassis to the 3rd pin earth has me kind of  . 

I don't see any conflict here.

I think I want the B+ center tap to be tied right at the first filter cap negative leg?

Yes, that's the idea. Do you mean the center tap goes to chassis, then  first filter cap negative leg goes to chassis via another connection? That is called "star-ground", which is a paramount concept in terms of safety, but not thebest idea in terms of noise performance, but acceptable on most valve equipment because currents are way smaller than in equivalent solid-state gear.

The IEC chassis ground on it's own lug to chassis near the IEC inlet. 

Yes, but you'll have to provide a connection between audio ground and chassis; this connection does not necessarily stiff - many designers favour the use of a low-value resistor shunted by a cap (typical 10-100r//0.1uF)

Still not sure what to do about ac heaters.

See above.

 

[MicDaddy]

Forgive me, some of this is probably only a lack of basic fundamentals on my part.

HT center tap is 0V, I was looking at it as if center tap was half of HT.  It makes sense treating it as 0V and tying to the negative side of the filter cap.

 

Yes, that's the idea. Do you mean the center tap goes to chassis, then  first filter cap negative leg goes to chassis via another connection? That is called "star-ground", which is a paramount concept in terms of safety, but not thebest idea in terms of noise performance, but acceptable on most valve equipment because currents are way smaller than in equivalent solid-state gear.

I was thinking to tie the HT center tap physically to the same node as the negative side of the first filter cap, in hopes to eliminate any unnecessary current path between the two.  The IEC earth would be on its own lug tied to chassis nearer the inlet. 

 

[abbey road d enfer]

I was thinking to tie the HT center tap physically to the same node as the negative side of the first filter cap, in hopes to eliminate any unnecessary current path between the two.

That is a very wise choice. 

The IEC earth would be on its own lug tied to chassis nearer the inlet.

Indeed. However, you have to think how you're going to connect "ground" to chassis. Some designers prefer joining at the input jack, some at the negative of smoothing cap, some at the cathodes of the output valves, some directly, others via RC ... many choices. Open to experimentation.

 

[PRR]

> When micing this amp, I get some good shocks

> installing a chassis to the 3rd pin earth has me kind of  . 

All touchable metal in the house MUST be wired back to the building Main Ground, which is (presumed) connected via dirt-rods to dirt (and concrete!).

Your 'lectric stove, dryer, metal outlet covers are all hard-wired via the green/bare wire to the fusebox ground.

Your guitar amp (where the user holds on to the ground which is invariably the metal chassis) *MUST* have its chassis tied through the 3rd pin to the power outlet.

This was not done through the 1960s. Your "old Airline amp" probably has the older 2-pin plug. This "can" be non-lethal. If everything is 100% fresh and tested, it isn't real dangerous at all. However all insulation will eventually fail, and that Airline has plenty of years and volt-hours on it. High time to install a 3-pin power cord.

That is totally apart from your inside wiring issues.

The Airline is a 9/10 copy of a Champ. Wire the main systems JUST like a modern Champ. Yes, the high-volt CT is surely grounded. The parallel heater system can be one-side-grounded (cheap, hummy), use a 6V CT, or use 2*100r CT. Any of these will work. The one-side-grounded trick fell out of favor in the 1950s (in guitar amps; still keen in FM tuners and some other uses). It might be worth the work to re-do that as 2-wire. There are LOTS of g-amp forums with pictures and advice.

 

[DerEber]

Push-pull pairs is precisely the case where wiring heaters in-phase is beneficial, because it is a common-mode signal, which is actually cancelled in the push-pull operation.

Yes..... you are right!!

 

[ruffrecords]

You may remember I was surprised at how low the hum level was with ac heaters in the tests I conducted on a Eurochannel mic pre. At the time I suggested that elevated heaters might have been a significant factor so today I changed the power supply so that the heaters were not elevated and repeated the test. Basically there was no difference between elevated heaters and heaters connected to 0V. The only proviso is that the heaters are balanced by a 500 ohm wire wound pot across them with the wiper connected to 0V. Adjusting the pot from one extreme to the other gives a null pretty much as the centre and a 3dB increase  in measured noise at the extremes. I repeated the test with elevated ac heaters and got identical results. As far as I could tell the null was at the same position in both cases and the increase in hum at the extremes was the same Here is a scope capture of the worst case noise and hum (max gain worst pot position). Adjusting the pot to the centre makes the hum waveform disappear completely.

Cheers

Ian

Edit: oops, forgot the picture!

 

[PRR]

> no difference between elevated heaters and heaters connected to 0V.

Elevation covers the case of heater-cathode leakage. Get it the right way, leakage is less. If the leakage is an unbalanced hum due to AC heat, then hum is less.

Apparently such leakage is NOT a common fault.

Golden-age designers elevated so they could use "every tube in the crate", even the few % with leakage.

Modern tubes may be better, or you just may be lucky in your small sample.

 

[ruffrecords]

Modern tubes may be better, or you just may be lucky in your small sample.

Good point. I am going to replace the pot with a couple of 100R resistors. Then I'll repeat the test with the original tubes and then with lots of others to see if there is much variation with tube sample.

Cheers

Ian

 

[ruffrecords]

I removed the hum dinger pot and replaced it with a pair of 220 ohm fixed resistors the centre of which was connected to the heater elevation point. No resistor selection, just straight out of the box. I then tested 15 different 12AX7 tubes. Some were EH,  some WA and some LPS variants. Every single one gave near identical performance. With gain at 53dB each and every one gave a an output noise level of -68dBu give or take a fraction of a dB. This makes the EIN -121dBu which is as good as it was with dc heaters.

I tried changing the output tubes (2 x 6922) but this made no difference. I also tried a couple of cheap Russian 6N23P equivalents to the 6922 and again the performance was identical.

Bear in mind this is an unscreened PCB lying on my bench. The only time I got a little more hum was when I switched on my soldering iron. This contains a big mains transformer and it was about 6 inches from the PCB and the mic input transformer. Even that only worsened the hum by 3dB

Cheers

Ian

 

[merlin]

With gain at 53dB each and every one gave a an output noise level of -68dBu give or take a fraction of a dB. This makes the EIN -121dBu which is as good as it was with dc heaters.

How are you measuring that? Is there an input transformer, and what was the input termination?

 

[ruffrecords]

With gain at 53dB each and every one gave a an output noise level of -68dBu give or take a fraction of a dB. This makes the EIN -121dBu which is as good as it was with dc heaters.

How are you measuring that? Is there an input transformer, and what was the input termination?

1:10 input transformer. Input shorted. Output measured with Lindos MS10.

Cheers

Ian

 

[PRR]

> I removed the hum dinger pot

If your heater wiring could be *perfectly* symmetrical to the key audio nodes, the "optimum" tap would be dead-center.

12AX7's awkward pinout foils this.

If you got dead-center (pot or 2 resistors), you have the knack.

When you don't, can't, or are not sure your factory-help can be so neat, you provide a humdinger (which used to be a trade-mark) to diddle with.

This also allows for a very leaky heater insulation, though the pot will probably have to be diddled daily until the tube just goes "bad".

Ampeg VT-40 with all heater wiring on the PCB: if the PCB held *only* heater wiring it could be symmetric, but Ampeg cluttered it with audio nodes and the heater wiring could-not be made symmetric to all nodes. At normal gain setings the humdinger optimum was consistently 45%, though this would change with some extreme settings. Today of course this problem can be solved, for production amps, with some relatively cheap diodes and caps to send DC to the heaters. Series-heat is not only more efficient (~~2V drop in Si diode bridge) but also also reduces the traces on the PCB.

Proper AC heat is, of course, "right" for old-school work.

> unscreened PCB lying on my bench.

Humdinger should not be used to correct for bad (or no) shielding. But it is interesting that you can get good results wide-open, and only mild hum from a too-close transformer.

 

[CJ]

you can also cancel hum in the output tube bias circuit (push-pull), maybe less likely to fry parts if done this way,


OT:
here is a  diagram for wiring mains, news to me on the fuse holder leads,

 

[merlin]

1:10 input transformer. Input shorted. Output measured with Lindos MS10.

Ah, that makes sense. It all seemed too good to be true until you said 1:10. That's 20dB of step-up even before you start! No wonder you can get away with AC heaters.  You dont' usually see such high step-up ratios in British designs (American, yes). I was expecting 6dB at most. All makes sense now.

 

[merlin]

here is a  diagram for wiring mains, news to me on the fuse holder leads,

You don't have to wire the fuse holder that way, it's just good practice since the fuse breaks contact with the live wire as soon as you start to pull it out.

 

[ruffrecords]

1:10 input transformer. Input shorted. Output measured with Lindos MS10.

Ah, that makes sense. It all seemed too good to be true until you said 1:10. That's 20dB of step-up even before you start! No wonder you can get away with AC heaters.  You dont' usually see such high step-up ratios in British designs (American, yes). I was expecting 6dB at most. All makes sense now.

1:10 is very common in tube designs. It is the only way to achieve a reasonable noise figure. (V72, REDD47 etc). 1:2 is common in semiconductor designs to match the optimum noise resistance of typical BJTs as I am sure you are aware.

Surprisingly, there is one well known British semiconductor design that uses a 1:10 input transformer - Helios.

Cheers

Ian

 

[merlin]

1:10 is very common in tube designs. It is the only way to achieve a reasonable noise figure.

OK, that makes sense. I am so used to thinking of regulated DC heaters -which leave you with the irreducible noise of the first valve- that it wouldn't occur to me to use more than 6dB of step-up. Not so for traditional AC-heater designs where you have hum to think about too.