Valve power supply - How to size transformer?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all,

I have the following valve line up:

2 x GZ34 (1.9A ea @ 5v)
2 x 6SL7GT (0.3A ea @ 6.3v, 2.3mA @ 250v)
4 x 807 (0.9A ea @ 6.3v, 140mA @ 400v)

So with a capacitor filter I need a transformer that can do the following:

5v AC RMS @ 3.8A Continious
6.3v AC RMS @ 4.2A Continious
290v AC RMS (just over 400v DC, unregulated) @ 565mA intermittently.

The B+ will only approach max very intermittently, but the heater supplies will be running constantly whether there is signal or not.

So do I give these specs to the transformer supplier or do I just order a thumbsuck like:

5v 4A
6.3v 4.5A
290v 600mA

Should I build in headroom for voltage sag under load conditions? I see these unregulated supplies everywhere, does anyone build a supply like this that is fully regulated?

Thanks!
Sander
 
Hi DF96,

One HT line, with the two GZ34's in parrallel to rectify. The application is a stereo audio amp that will in all likelyhood never be driven hard, just playing at living room levels. The design intent was to make it glow, not to raise the roof...

So I don't know about duty cycle, heaters 100%, B+ also 100% while it is on, but not at max levels. Amp will have push pull on each channel, so mode AB1.

Actually, being push pull, both tubes will never have full B+ simultaneously, so each has a duty cycle of 50%, or the two together has a 100% duty cycle, but only at 140mA not at 280 like I originally thought? Unless I am totally missing the point.

The B+ is also a centre tapped winding, so I assume 300-0-300, or 290-0-290... How important is 10 or 20 volts either way on something like this?

Thanks!
Sander
 
OK. Was 140mA for one channel or two? In normal hi-fi use the DC current consumption won't rise very much, because signal peaks are only present for a small fraction of the total time.

Your best bet is to use PSUD2 to model the HT supply. This will tell you what the secondary RMS current will be. Alternatively, just guess that it will be 3-4 times the DC current.

10 or 20 volts either way will not matter.
 
.
Actually, being push pull, both tubes will never have full B+ simultaneously, so each has a duty cycle of 50%, or the two together has a 100% duty cycle, but only at 140mA not at 280 like I originally thought? Unless I am totally missing the point.

Yes. correct but you want a power transformer that has a max AC current rating that is at least 2X the max required DC current. This is provide a certain amount of load regulation.

Some will say to go even larger than 2x but if the intent is low "bedroom" level sound levels you don't need the headroom. Go larger than 2X if you intend to fill up a 20 foot square room with normal music volume levels.

You are right that a few volts one way or the other on the B+ does not mater to much. The way to think about it is in "percent". Certainly a 5% difference in B+ does not matter much. However I've played around with a Variac transformer where I can adjust the B+ as I listen and mostly "more is better" as long as do don't go past limits.
 
Thanks Chris,

This is the 'rule of thumb' advice that I was looking for. I was just concerned that if I build the transformer to specs it may turn out to be too small. Does the same 2x approach also hold for the heater supplies? Being constant draw I assume it is easier to design.

I have also thought of the variac approach, but that would imply splitting the 5v & 6.3v transformer from the B+ transformer? What is the normal practice here?

Also if I only need 140mA per channel, I should be OK with just one GZ34 rectifier instead of two in parallel. This is a circuit for an old commercial amp that I was trying to reproduce and they had two GZ37's in parallel. I have already changed that to GZ34's as they are more readily available and the specs seemed fine.

DF96, I have downloaded PSUD2 and am trying to model my power supply, but this will apparantly take a few extra days to figure it out! :)

Thanks!
Sander
 
I have also thought of the variac approach, but that would imply splitting the 5v & 6.3v transformer from the B+ transformer? What is the normal practice here?

On consumer equipment heaters are usually supplied from the HV transformer.
On commercial and industrial equipment the heaters are usually supplied from their own power supplies. Heater voltage and correct application time are critical in obtaining good tube life so applications with expensive tubes can afford some increased power supply complexity, most tubes at work turn on the heaters 10 minutes before the HV. All heater supplies have adjustable voltage.

If you are building from scratch then there is little reason to use the same transformer for heaters and HV, this was done in the consumer equipment of yesteryear for economy but it might work out cheaper to use separate transformers these days due to availability
 
The Tubes have large indirectly heated Oxide cathodes and ion pump getters, enabling the tubes to be held on standby with reduced heater voltage for long periods of time. Tube life 10k - 50k hrs depending on the device. Thermal cycling and arcover are the main enemies of high power tubes, loss of cathode emission kills them all eventually.

I am not suggesting 10 minute warm up for audio tubes, just enough time to get some cathode emission before applying the HV, this is mainly to reduce any chance of arcover. If the equipment uses a choke input filter, the HV is going to go sky high before the output devices start drawing current. Older equipment had the delay provided by the thermionic rectifier which doubled as a soft starter not that it was good for rectifier life. Mercury vapour rectifiers required heater power first.

I may be a bit rusty on the small stuff AFAIR it was always preferable to have heater power then HV not the other way around also control grid bias needs to be applied before HV.

Other advantages of separate supplies is the ability to soft start the heaters (reduced thermal shock) and apply correct voltage independent of any HV load, maybe this is overkill for audio where the tubes are cheap. The Older ARRL handbooks and the RCA transmitting tube manual are good references.
 
High power (e.g. transmitters) usually means high voltage and then the advice about delaying HT is valid. For normal domestic audio voltages arcing is unlikely. A choke input PSU will only go 40-50% over normal, so provided the capacitors can cope the valves will probably be OK. If not, a string of zeners can waste a bit of power until the valves warm up.

Valves which are intended to run for significant periods with full heaters and no current (e.g. early computers and other logic) have special cathodes which do not form interface.
 
Going a bit OT
The cathode interface phenomenon was new to me so I did some digging It turned out to be a chemical reaction (reduction) between a cathode substrate constituent and the oxide coating, as a result it is an integral of a time temperature function. Meaning that it is mainly dependant on heater on time and temperature. The same factors govern loss of emission.

There may be some electrochemical process as a result of cathode current which might partially inhibit the reaction due to the cathode coating being positive with respect to the substrate during current flow favouring oxidation not reduction.

It is unlikely a short heater preheat for an amplifier is which is going to be run for hours would have much effect. It can be avoided by choice of a suitable cathode substrate or barrier layer, whether or not a particular tube manufacturer does so would have to be gleaned from their application notes. It only affects oxide coated cathodes so 833A users have no need to worry :).
 
Hi guys,

Thanks for all the advice! And one more question to prove my ignorance:

Keeping to round numbers: A secondary 0-100V @ 1A through a solidstate bridge rectifier gives approx 140V DC @ 1A.

And a secondary 100 - 0 - 100 through 2 diodes / valve rectifiers also gives 140V DC @ 1A.

So when calculating the seconday VA on a centre tapped secondary you obviously can't work on a 200V swing @ 1A. Do you just always use half the total voltage for the VA calculations?

Intuitively the power drawn is the same between the two transformers, does this mean that half the voltage is 'wasted'?

And then I assume you can run a 'normal' 0 - 100V transformer with 4 valves to maky up a bridge rectifier, has anyone ever done this?

thanks!
Sander
 
Hi guys,

Thanks for all the advice! And one more question to prove my ignorance:

So when calculating the seconday VA on a centre tapped secondary you obviously can't work on a 200V swing @ 1A. Do you just always use half the total voltage for the VA calculations?
No there is a utilisation factor which depends on the rectifier and filter setup. A bridge rectifier with a choke input filter has a utilisation factor near 1, so the secondary VA rating is similar to the load power, capacitor input filters have utilisation factor between 1.5 and 1.8, full wave centre tapped rectifiers with choke input filter are around 1.5.
Intuitively the power drawn is the same between the two transformers, does this mean that half the voltage is 'wasted'?
Not quite but it does mean that the secondary needs more copper than with a bridge rectifier because current is only flowing half the time so the winding can be made roughly 0.7 times the area of a winding in which the current flows all the time but there are 2 windings so there is 1.4 times as much copper in the secondary.

And then I assume you can run a 'normal' 0 - 100V transformer with 4 valves to maky up a bridge rectifier, has anyone ever done this?
thanks!
Sander
Yes but the voltage drop across thermionic rectifiers is quite high, even mercury vapour ones have 15V drop, I have seen one mercury vapour bridge rectifier but you need 3 heater windings for a single phase and 4 for a 3 phase, bridge rectifier.
 
One more thought, If a toroidal HV transformer is available, it is a trivial exercise to wind a few turns of hookup wire around the core to get one or more 5V windings for the rectifier heaters. Saving the expense and bulk of another transformer. Because the primary will be oversized to get the appropriate secondary rating there should be enough surplus capacity to run 20 or so watts of heater.

If you get a custom transformer made then the transformer company will know what size secondary wire to use if you tell them the rectifier type and filter capacitor size. The last time I had a custom transformer made it was reasonably priced YMMV.
 
Thanks guys,

I will have a custom transformer made, but I have been playing with adding windings onto toroidals just to test out concepts. Toroids really make it easy to play with, but they don't quite look the part in a valve amplifier! :)

I know with normal EI transformers good practice is to turn them so that the magnetic fields cut each other at right angles.

But if I have more than one toroidal transformer, can I stack them on top of each other without causing a problem? I assume I can't have a conductive bolt through all of them as that would constitute a one turn winding through all the transformers, but with a non-conductive central shaft this should not be an issue?
 
Max anode dissipation for an 807 (it's just a 6L6 after all), is 25W. That means max dissipation for 4 off will be 100W, at 400V is 250mA max. Frankly its a bit much and I wouldn't run them at more than 20W each so 200 mA for the anodes is tops. Allowing 10 to 20 mA for the screens and a tiny bit for the rest of the circuit means a 250mA capable supply is more than adequate.
Regards
Henry
 
valve power supply

Can anybody point me to some free source "how to design power suppies for valves".
Ive been looking through some schematics, and it looks like after the rectifier there are resistors in series giving several HT in diminishing order, as each HT "point" is capped to ground.
Usually the voltage goes down as the resistors go on,
the latest valve stage getting the higher voltage.
I see some pattern, but how to calculate this?
View image: power supply
For example, I am going to use cheap transformer scheme (220:6v - 6:220v, 2 transformers and getting the heating in the middle), so I need to calculate my power supply for 220 volts raw.
Also, I newer managed to understand how exactly the voltage changes after diode bridge and capacitors - seems to go higher, so I don't know which voltage to count on - the transformer's raw, or whatever I measure after rectification/capping???
It is best, If I understand how to calculate it
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.