807 amp psu

[adibrook]

I'm still putting together ideas for an interasting stage amp.

A big possibility is to use 807's for the output stage. After thinking about it for a while, i decided that although 807's arnt ideal, they will be ok.

813's are better, but working with KV's might get abit expensive and could hurt abit.

I have a pretty big transformer, which fetured in other threads, whoch is kind of suitable. However, it has no 6.3V winding, so i also have a 6.3v heater transformer.

So, i'v attached several possible designs i'v thought of.

The 1st one is just a normal setup, like you would have in a radio or a small amp.

That was the original psu for the 1st 807 amp prototype. It worked, but i never had the chance to push it hard.

The problem with that is that whereas thers plenty of current, voltage is abit low really. I think maybe 807s need slightly more anode voltage.

So, then thers circuit 2. This will rectify the full 600v winding, which would be enough to run the 807 to its limits. However, then the current becomes abit of an issue.

Ok...3rd circuit.

This would involve me finding a isolation transformer. I put down 200VA as a rough guess to keep it below limits. Its a pretty common size as far as i know.

That thing on the secondery is what i belive to be a doubler. I found one in my TEK scope. As far as i understand in the scope it turns 190vRMS into 500VDC. So it it shoudl manage quiet alot of volts even under heavy load from 240v.

And best of all...current is only limited by the diodes, and the transformer, which is big anyway. I could use a quad of 807's, and then their filiments can run from the 25.2V winding. That would be quiet a large amp.

Also maybe i can run them with fixed bias nearing classB, which would need a negative voltage source...(not a problem...thers transformers all over the place anyway one more wouldnt hurt), but that would be pretty powerfull.

any ideas?

Thanks...

[Skywave]

Hi Adi.

Re: cct. 3. It's a half-bridge rectifier: the caps./ resistors are "replacements" for the "missing valves".

Another way of regarding it is is that it is two half-wave rectifiers with their O/Ps combined in series. And that's the snag. It isn't a very effecient rectifier scheme. You'll need big caps. if you are going to draw any appreciable current and keep the 50 Hz ripple down. Plus the "top" valve has its heater "well up in the air": so a special, high-voltage insulated winding for its heater is required.
You don't specify the secy. voltage of the isol. transformer. If it's 240v, then the peak volts across each cap. will be 1.4 times 240 = 336 v. Two caps. in series = 670 volts approx. on no-load. For a 200VA transformer, that's about 750mA avail. from the sec'y., but after about 100mA the d.c. voltage with this config. will probably start to drop off quite quickly - with a lot of ripple on the HT. And as I said, you'll need big caps: something like 1000uF (min.) at 400v dc each, as a quick, estimated guess.

But like most things like this, these comments are based only on "rules of thumb". IMHO "Rules of Thumb" just stop you from trying to do the impossible! Or at least, save you from wasting your time! There really is no substitute for doing a lash-up and experimenting.

Experimenting: the sort of stuff that you - and I - really enjoy!

Best wishes.

/ Al.

[adibrook]

Thanks for that explanation. I never got how that circuit worked before.

In the TEK scope the entire power of the whole scope goes through one of these, and the caps are 125uf with 150k resistors. But the diodes are silicone.

The scope obviously doesnt draw as much current as quad 807 power amp.

Allmost every TV i'v taken appart has a huge cap of around 320uf 400v in the psu. I'm building up quiet a nice collection of them. Since they are used in allmost every TV, than they must be produced alot and easy to get.

I dont know what values they go up to, but i know from rippign appart tv's that 320uf 400v is a common value.

Say if i used 4 of them in paralell for each ''leg'' of the doubler. That would give 1280uf each. That should work.
As far as i know p/p output stages are better at ripple rejection than s/e anyway.

So you think this is a realistic circuit for practical use if i find big enough caps ?

BTW by isolation transformer i mean those thigns used for troubleshooting dodgy 240v appliences. Just a 1:1 240v transformer.

[Skywave]

Hi Adi,
I need to add a few corrections to my earlier Post - which was written late & I was pretty tired!

The "official" name for cct. 3 is a full-wave voltage doubler. Therefore the ripple is at 100 Hz, not 50 Hz. Therefore, the smoothing requirements are not so severe. For up to about 100 mA drain, each cap. at 320 uF / 400v should be adequate - but it's always a good idea to leave enough space on the chassis to add two more - if this proves necessary on test. There is always the option, of course, of adding a smoothing choke.
On no-load, the O/P voltage will be about 700v., dropping quickly to about 480v. when a current of about 100 mA is drawn.

The only two snags I can see with this config. - but which can be overcome - are that:
1. The neg. pole - and therefore the outer case - of the "top" cap is at about 240v dc (min.) above 0v. dc ;
2. The heaters of both rectifiers are well above 0v dc, and each is at a different potential to 0v dc. To avoid the risk of heater-cathode insul. failure on either valve, each rect. needs to have its own high-insulated winding.

This latter problem, of course, doesn't arise with solid-state rectifiers - a la, the TEK 'scope.

So, with these corrections, sounds like a "go-er" to me!

/ Al.

[Skywave]

Just had a further thought on cct. 3 which may be of interest.

I'm sure that you are familiar with the older type of PC PSU where there is a switch at the back where you can configure the PSU for 230 v ac or 115 v ac input.
They way this is done is to use cct. 3 for 115v ac I/P
For a 240 v ac I/P, a full-wave bridge rect. is used, charging two caps. in parallel. When the switch is set at 115v ac, the switch reconfigures the bridge and the caps. to your cct. 3. With this mode, since there is only half the ac input, you need a voltage doubler. And with the switch in this "115v ac" position, that's exactly what you get.

Clever, innit?

Al.

[adibrook]

Ok. Thanks.

I think heater supplies for the rectifiers are no problem. I know that some transformers for 300B amps have two isolated 5v windings for example, and are not very expensive.

Or just a couple of bog standard 5V transformers, liek the ones used in small wall warts and power supplies for small electronics. Theyre pretty common and cheap. The insulation between primery and secondery should be able to stand at least 240v fine, since they are used for mains.

The cap issue isnt a veyr big deal i dont think...modern caps dont use the can as a connection anyway, so no insulating sleeve (like in the tek) is neseserry.

What are the 150K resistors for in the tek? Are they bleeders to stop the caps storing charge? Or do they have another purpurse? What value shoudl i consider using?

Thanks.

PS

I will rip open my old computer psu thats been in my drawer for ages and have a look.

[Skywave]

Hi again!

The Rs. across the caps. are there for 3 reasons:
To equalise the voltage across the caps.,
To act as bleeders / dischargers for the caps.,
To provide a minimum load to stop the O/P voltage from soaring too high on no-load / reduced load. Kinda helps with the regulation - in a crude way.

With this in mind, a little calculation gives:
For 500v dc on load, 250v across each cap. Let's say we are drawing 100 mA dc O/P. A good rule of thumb is to choose one-tenth of this as a bleed current: 10 mA. So, 250v across a R that is passing 10 mA (common current in both Rs) gives 25 K ohms. It will dissipate (250 * 0.01) watts = 2.5 watts.
Allow for operation under no-load conditions & let the Rs have a "comfy" life: choose 5w rated Rs. I suggest wire-wounds for reliability.

The value of 25k isn't critical. If, on test, you find that there's a little too much ripple, reduce the current through these Rs so that there is less of a drain on the transformer / rectifiers / smoothing caps. Try 50k each. I wouldn't go above that value, though.

Just a further thought. In cct. 3, there's no reason why you can't add another single cap across the final dc O/P. Just need to make sure that it is of adequate working voltage.

And another. Make sure that the voltage at the rectifier heater valve pins is correct. Too low a voltage at the heater pins is a frequent cause of valve rectifier failure.

And with that - I'm off to bed!
'Night!

Al.

[boiss]

Hi Adi,
Once I get my heater supply sorted and have rewound my toroid (O/P), I'll get on with my 813 pp amp, I've got an 800V 400VA transformer for the Ht supply, but it's gonna get silicon for rectification.
Mike.

[Top Cap]

Hi Adi, I think you have the reccy's round the wrong way, should have their cathodes to the positive output rail. Don't forget to add 10% to your transformer outputs to cater for mains variation and watch the piv ratings of the rectifiers. The 5R4 looks ok for this application as it has a piv rating of 2800V whilst able to take 1kV on the anodes with a current capacity of 150mA. However the maximum resovoir capacitance is quoted at just 4uF unless you add some series resistance.
Les

[Skywave]

Originally Posted by Top Cap

Hi Adi, I think you have the reccy's round the wrong way, should have their cathodes to the positive output rail.

I can confirm that - in ccts. 2 and 3, the +ve O/P should be taken from the cathodes.
Question is - how come I missed that?

Al.

[Merlin]

Originally Posted by adibrook

Ok. Thanks.

The cap issue isnt a veyr big deal i dont think...modern caps dont use the can as a connection anyway, so no insulating sleeve (like in the tek) is neseserry.

Modern caps still have the casing connected to the negative terminal, so don't let it touch anything else.

[boiss]

Adi,
There was a post not so long back which had the caps mounted between two plates of perspex, admittedly they had screw terminals but the basic idea is good. You could have drilled holes where the terminals exit.
Mike.