Isolation Transformer question..

[chptunes]

Hey folks.. i'm considering the Kay model 703 amp, for my next amp project.. (with some updates and changes)

I think the Hammond 187C120 Isolation Transformer would be a good fit.. it has a center-tapped Secondary, which allows Full Wave Rectification, and it's rated for 100 mA of current.

Model 703 schematic:  http://img407.imageshack.us/img407/2723/kay703dw1.jpg

187C120 data:  http://www.hammondmfg.com/186_187.htm

What do you guys think?  I can obtain the 187C120 for around $12.  Good choice?

[R.G.]

I think the Hammond 187C120 Isolation Transformer would be a good fit.. it has a center-tapped Secondary, which allows Full Wave Rectification, and it's rated for 100 mA of current.

A couple of things come up in my mind.

First, your rectifier, the 35Z5 can't do full wave rectification. It's a single diode. So unless you're replacing it with something else, full wave won't help. And if you did go to a dual rectifier, you'd need 230Vac CT to keep the secondary B+ at the same voltage.  So I don't think the 120Vct helps you. It doesn't hurt you necessarily if you just use it for isolating. Just saying.

Second, one of the things that you need to accomplish is to get the heater string off the primary AC and onto an isolated 115 secondary. This tranny has two 115V primaries and one 120Vct secondary. You could run it backwards to get 120Vac in  and two isolated 115Vac out, but the voltages will be smaller, maybe 105-110Vac on the perverted primaries, because of the way the transformer has to be wound for voltage sag under load.

Third, the 703 puts the heaters on the primary side, and it's B+ transformer only runs the high voltage. You need to worry about the heater string current added to the B+ current for how much current your isolation transformer provides. The heaters are 150ma on the tube string, so your transformer needs to be rated at well over 100ma, probably more like 250ma.

I recommend that you get something like the Hammond 186D120 or 187D120 unless you can measure the actual AC current under full-power operation and find it's smaller than the 250ma these provide. And I recommend that you simply put this whole transformer in front of the whole amp, not dinking with different secondaries and such. It's a simpler job to do electrically.

As always do not try this stuff unless you know how to wire high voltage AC power safely. We do not want to lose you or someone else to a wiring mistake.

[chptunes]

Thank you, for the valuable input.  I was just planning to use two Rectifiers for Full Wave, while grounding the Center-Tap.

Your thoughts open up a bunch of new questions for me.

I hadn't considered the Heater current, for the Transformer ratings.  The 250 mA units are relatively affordable/available too.

I'll sketch and post my idea for this power suppy.. ..see if we can get down to the nuts and bolts.  [I love this stuff]

[R.G.]

Thank you, for the valuable input.  I was just planning to use two Rectifiers for Full Wave, while grounding the Center-Tap.

The biggest difference that half-wave versus full-wave rectification makes is that half-wave needs twice the capacitance in the filter caps to smooth it to the same degree (more or less).

If you go full wave, you have to decide whether to do SS, tube, or a combination. SS will up the B+ voltage by about 40-50Vdc. Another tube would keep the B+ about the same, but then you will have to modify the heater string to supply two rectifiers. The tubes you have now are "matched" in that they need 150ma at a voltage matching the first numbers in the tube type; that is, the 35Z5 needs 35V at 150ma, and the 50L6 needs 50V at 150ma. The funny tube numbers belong to a set of tubes designed to have their filaments add up to 115Vac for home radios to save money on filament transformers. The "American five-tube set" is one such. They all need the same current, so they can be hooked in series to get to the AC voltage and all is well. The Kay uses two of these tubes, the 35- and 50- and then the 12AU6, which only needs 12V, so the heater voltages add up to 97Vac. There is a resistor in series to soak up the other 17V from a 115Vac line.

If you go full wave, you'll have to decide whether you put another 35Z5 in its own parallel string, as it can't go in series with the others, there's not enough voltage left over. It can't go in parallel, because the strong only conducts 150ma, so either the rectifiers would be starved for heater current or the other tubes would be cooked.

On solution you might consider if you must have full wave instead of a bigger filter cap is to use two solid state diodes to make full wave, then run THAT through the 35Z5 to drop off the extra high B+. This is harder on the 35Z5, as it's conducting twice the current now, but at least the heaters would be happier.

[I love this stuff]

Me too. 

[chptunes]

how bout this:


windows 7 screen shot

[R.G.]

If you add up the heater voltages you have, you get 3*35=105 plus the 12V for 117. OK, that works in terms of the heater voltages totaled. I'm guessing you're using the 35C5 instead of the 50L6 for an output amplifier tube. Is that right?

There are two issues with what you've drawn.

The biggest problem is that the heaters are still on the primary side of the transformer. Heaters have maybe 300V of isolation from the cathode, max. Modern safety standards started with 1.5kV, then 2400, then today's 4kV isolation from primary conductors. A spike on the AC line will puncture the oxide insulation on the heaters and punch through to the cathodes, possibly destroying the tube, and possibly electrocuting YOU if you happen to be standing in the wrong place or kissing a microphone, touching a mike stand etc. The heaters have to go on the secondary side of the power transformer for safety reasons.

The other problem is that the amp won't work with the B+ that comes out. (I view this as a smaller problem than having a reasonable chance of dying.)

The 187D120 puts out 120V center tapped, so the voltages to each side of the center tap are 60Vac. Each half-wave rectifier gives you only half of the voltage of the original setup, if I read the original schemo right. the original would make something like 160-170Vdc before sag. Your setup will give you only half of that, so the tubes will be starved for B+ voltage. It may work, but not like it originally did. To make this setup work in terms of voltages, you'd need a 240Vac secondary with a center tap.

Here are some options.
1. Use the 187D120 and ignore the center tap, using only the full 120Vac winding. Then
    1a. connect the existing power transformer to the 120Vac secondary, and play in happiness. The amp is a little heavier, but safe and little work had to be done.
    1b. remove the existing power transformer and wire in the 187D120 as follows:
         - use the fusing and AC switching as shown; worry about where your power-on indicator light gets wired, if the amp uses one
         - move the entire heater chain, including resistor, from the primary side to across the 120Vac secondary
         - put in a 600V to 1000V full wave bridge rectifier; this could be an integrated bridge, or something like four 1N4007s
         - run the (-) side of the bridge to the (-) side of the first filter cap; run the (+) side of the rectifier bridge to the anode of the 35Z5, then the cathode of the 35Z5 to the (+) side of the first filter cap
     1c. do 1a, but double the value of the first filter cap; since you're already replacing all the electrolytic caps with new ones (you are, right?) simply doubling the first filter cap gives you the same ripple reduction as going to full wave; you can only do this if the new, doubled capacitance is within the maximum capacitance spec of the 35Z5, of course.
2. Find a 240Vac CT transformer;  move the heaters to one half of the secondary, and use two SS diodes to full wave rectify, and the 35Z5 to tube-ify it.
3. Find a transformer with THREE 120VAC windings, and put the heater on one of them, using the other two as a full wave center tap. This is likely to be the quietest setup if you then make a fake center tap with two equal resistors across the heater winding and ground that fake center tap.

There are others, but they get more complex and change the amp more. Number 1a is the closest to how it is now.

[PRR]

You are over-thinking this.

The 5-toobe AC/DC radios (and their cheap-amp offspring) are highly integrated. Once you start modifying, you are on a slippery slope to different sounds, different tubes, *6V* tubes, and ultimately loss of all the funky goodness that these amps sometimes give.

The half-wave supply is part of the "charm". It is not a problem: the double-big filter cap is cheaper than extra windings and rectifier. (Yes, 1N4007 is cheaper, but you MUST drop many Volts in the heater-string and the 35Z5 is the obvious path.)

Leave the Kay AS-IS.

Make a 117V "wall outlet" with NO copper connection to the wall. An Isolation Transformer. Put it in a box in the speaker area. Hard-wire the Kay's "117VAC" leads to this IsoFormer. (I assume the Kay's power cord is old and very dangerous.)

Use 3-pin cord from real wall to IsoFormer. The Green lead must bond to box (if metal), IsoFormer frame, and then to a SOLID bond on the Kay's chassis.

A suitable IsoFormer is:

Transformer; Isolation; Pri:115 to 230VAC, Sec:115VAC; 50/60Hz; 50VA; 1500Vrms
Mfr. Part#:N-68X
Allied Stock#:70218526

[R.G.]

Yep. That is option 1a.

[chptunes]

This will be a scratch build.. I don't have a Kay amp.  I'm free to build/change/adjust however needed.

So, the 186D120, wired backwards, with two "perverted" Secondaries sounds like a cool option.  Heaters on one Secondary.. Half Wave Rectifier and B+ on the other.

• I wonder about the B+ Voltages with this option..
• I need to find the value for R1..

What do you guys think about this?


screen shot windows 7

[R.G.]

So, the 186D120, wired backwards, with two "perverted" Secondaries sounds like a cool option.  Heaters on one Secondary.. Half Wave Rectifier and B+ on the other.
• I wonder about the B+ Voltages with this option..
• I need to find the value for R1..
What do you guys think about this?

1) it's OK-ish in terms of operation. Since transformers are designed to sag to the rated voltage on the secondary at the full rated current, the no-load voltage on the secondary is actually higher than the rated voltage.  This is noted in the data sheets as the "regulation" percentage, which is a misnomer for advertising perposes; it's actually the amount of sag from no-load to full load. In a 30VA transformer, this might be 10% sag. So the secondary voltage is the voltage at full load, and this voltage rises to the regulation % HIGHER at no load.

If you feed the transformer backwards, the "120Vac" secondary is actually maybe a 132V secondary, and this makes the transformer ratio 132 to 115 each. So if you then feed 115 into the nominal secondary, you get out of each of the nominal primaries something like 115 * 115/132 = 100Vac. The prep work the maker did to get you high enough secondary voltage works against you if you work it backwards. This may be fine, as your B+ might come out at 130-140; low but usable.

2) The safety isolation from primary voltages to secondary voltages is still OK from winding to winding, but the details of how the secondary is wound versus the primary may make the isolation from the 120V secondary being used as a primary less effective to the core. Probably OK, but that may not have been looked at in the safety reviews. I would not use it this way unless the core was bolted to a metal chassis which is third-wire-grounded solidly.

3) floating the heaters is probably quieter, OK, but they still need to be tied to signal ground somewhere. It's best if you can float the heaters and pick a point that's in the middle, and ground that. You can either use two equal-value resistors that don't eat a huge amount of current across the whole heater winding, and ground the middle, or try to arrange the heaters and the resistance(s) in series to make a "CT" point out of the heaters themselves. You're not going to get perfect balance on the heaters, but it'll be better in terms of hum than only grounding one end.

4) Resistor value is easy. The heaters on the tubes want the rated voltage of their first type number (i.e. 50, 35, 12...) at 150ma, and they were picked to make that true. So whatever voltage you feed the series heater string, R1 is the resistance that makes 150ma come true when those voltages are subtracted from it.

Example: The winding feeding the heaters puts out 100Vac. The heaters need a total of 97V. This leaves 3V for the heaters. R1 is 3V/0.15 = 20 ohms, and the power is 3V*0.15 = 0.45W, so a 1W resistor is fine. I would be sorely tempted to leave R1 out (i.e. R1 = 0 ohms) at this very low voltage across it.

Example: The winding feeding the heaters puts out 115Vac. The heaters eat 97 of that, and leave 18V for R1. R1 has a voltage of 18V and a current of 0.15A, so R1 must be 120 ohms, and needs a power rating of over 18V *0.15A = 2.7W. 3W is marginal, I'd use a 5W wirewound.

[chptunes]

So, omit R1.. find a "Center Tap" to Ground the Heater String..

I've read a bunch of tube Datasheets, and Merlin's "How To.." website.  Apparently, most amp manufacturers have been neglecting the inclusion of a 'Plate Resistor' on their Rectifier Tubes for decades.  I wonder if I need R2, in this circuit...

http://www.valvewizard.co.uk/

How about this?


screen grab

[chptunes]

..for slightly less output, all three tubes could be minature button 7-pin based.. 35W4, 50C5, and 12AU6.. like this old Gregory Mark I schematic:


image hosting 12mb

[R.G.]

So, omit R1.. find a "Center Tap" to Ground the Heater String..

That's where I'd start.

I've read a bunch of tube Datasheets, and Merlin's "How To.." website.  Apparently, most amp manufacturers have been neglecting the inclusion of a 'Plate Resistor' on their Rectifier Tubes for decades.  I wonder if I need R2, in this circuit...

Depends. They don't specify a plate resistor, but they do specify a maximum value of the first filter cap.
Um, I went off to look that up in my trusty Sylvania tube book and the *DO* specify a plate resistance as well as a maximum cap value!
For the 35W4 and the 35Z5, these are both "minimum plate circuit resistance = 15 ohms" and "first filter cap maximum 40uF". There's some other nonsense in there for panel lamps in series and such, but that's for the five-tube radio set, I think.

So yeah, there is a plate resistor specified, and it can be in the transformer secondary, or it can be a separate resistor, like R2. The issue is that if peak currents are too big, they eat the emitting oxide off the cathode of the rectifier tube. Big caps and low resistances equal high currents, so they tell you in effect, how to limit these. Most rectifier tubes only say a maximum first filter cap.

How about this?

I'd leave R1, put another "R1" on the other side with the 35 and the 12, and adjust the two resistors for lowest hum. I'd leave R2 in there as 12-15 ohms and make the first filter cap 39uF *max*.

[chptunes]

thanks a bunch RG and PRR.. I may post a full schematic, based on this power supply.

[PRR]

> I don't have a Kay amp.  I'm free to build/change/adjust however needed.

So where are you going to get the cheap-crap output transformers KAY got from radio rejects?

A more-rational build would use 6Y6. This "is" 50L6 but with 6V heater. 6AU6 is the obvious preamp/driver tube. The 35Z5 adds nothing; use 120V to FWB to 165V DC, then a dropping resistor and 2nd filter cap to kill buzz and get down into the 105V-125V DC range that radio outputs were intended for. Or drop minimally to 150v and re-adjust 6Y6 current and impedance for somewhat more power than the AC/DC form got.

> find a "Center Tap" to Ground the Heater String..

No, to Center the AC to the First tube. V1 has two heater pins. The voltage across them "should" be equal-but-opposite so the 60Hz electric field around the other (audio) pins almost-cancels.

Since the first tube has gain of 50, this is 50 times less important at the 50L6.

There's ways to optimize the first tube in a series-heater rig, but if you really care, THE way is to run a lower-voltage (usually parallel heaters) and CT that.

> they *DO* specify a plate resistance as well as a maximum cap value!

It is usually a "typical cap value" and the resistance needed to protect the rectifier.

As I don't think 35Z5 has "a sound" (A class-A amp has nearly constant current draw, no "sag"), it just hastens the heat-death of the universe. Use sand-diode.

[chptunes]

This is where I was headed:


free upload

The 6Y6 tube that you mention will produce over 3 Watts with 135V on the Plate!  And the 6.3V Heater draws 1.25 Amperes!

I'd need to rethink the Power Supply for such a design..

[tubegeek]

This will be a scratch build.. I don't have a Kay amp.

Then, pardon my bluntness, you're insane.

Why in the name of DeForest would you even contemplate a heater-string design? It's not like that Kay amp is some sort of widely-admired paragon of tone or anything, nor is it a unique enough design to check out just for the quirk of it. Meanwhile you are committing your time and attention to an amp that you won't be able to change (I bet you will want to) and a power supply that won't allow you to do anything else. Unless you've got a dozen homebrew guitar amps you've already built, sitting in your closet, this is a poor choice. If you are just starting out, I'd STRONGLY recommend building a design that uses common-but-cheap tubes, standard heater voltages (so 6 or 12) and either nine-pin or octal preamp tubes and octal power tube(s).

Unless I'm missing something - do you have a stash of heater-string tubes you are trying to build with? Or is there some virtue to the Kay amp that eludes me completely? I don't mean to pee in your cornflakes but I would love to see you have a much more open upgrade/experimentation path available to you once you finish your build.

If you decide you want more gain, you switch to a dual-triode preamp tube. Or punch another small socket. If you decide you want to up the power, this can be done fairly easily by an output tube swap and/or raising the B+ by reducing the dropping resistor PRR mentions in the power supply filter. If you decide you want a tone stack or a tremolo... you see where I'm going with this.

Look, I'm all for building crazy/weird/cheap garbage, I've done it a lot. But I did it when I was pulling parts out of discarded tube TV's and radios that I was paying zero for. Not when I was starting from scratch and sourcing all the parts at retail. If you're going to put down money to buy all the bits and pieces, get better value than a soon-to-be-useless 50C5. Hammond 125BSE is a fine choice for your output transformer so I'd stick with that, just know that you won't get the "my-transformer-crapped-out-two-watts-ago" distortion from it, you''ll get decent fullish bandwidth and reasonably full rated power output.

[PRR]

> The 6Y6 tube that you mention will produce over 3 Watts with 135V on the Plate!  And the 6.3V Heater draws 1.25 Amperes!

The 50L6 heater needs fifty Volts!!

6Y6 and 50L6 are, for any practical purpose, "the same tube".

Or perhaps: 6Y6 is *marginally* beefier (also taller).

50.0V at 0.15A is 7.5 Watts.
6.3V at 1.25A is 8.875 Watts.
(Not exact the same, but there's some rounding.)

50L6 is specified for applications working without a Power Transformer. In vacuum-rectifier days, that meant B+ on the order of 110VDC. If you have 6.3V heat, you can probably pick any B+ you like, so the 6Y6 opts to show 135VDC data. Both also have a 200VDC op-point. The 6Y6 shows a higher number, but the 50L6 is shown somewhat under its dissipation rating.

Curves are suspiciously similar.

I really think the 50L6 and kin are the low-cost radio tubes, and 6Y6 is for designers who want "the same thing" but with a power transformer (and the more generous budget/expectations which come with that).

FWIW: I have a Kent amplifier of similar plan. It NEEDS an iso-former, and the usual re-capping (and the stock speaker is way-too-lame), and I don't have a real use for it, so it sits. Maybe we should do a deal.

[chptunes]

Then, pardon my bluntness, you're insane.

The 50L6 heater needs fifty Volts!!

Hey.. if I didn't know any better, I'd think you guys were trying to deter me.



OK, OK.. I got to this point because I thought I'd see how few parts and dollars are required to build a tube amp.  Power Transformers are expensive.

I'll keep this 50C5 design, and work up a 6Y6 design to compare.

..thanks for the motivation...

[chptunes]

{by the way, there are internet shops that have gobs of these tubes (New Old Stock) for less than $10 each.}