900v B+ Silly time.

[Layberinthius]

I'm considering a 200mA 900v un-regulated (but well filtered)
power supply for a combined SE 813 Class-A without a seperated power supply chassis, using:

Sowter M045 Centre Tapped.

HT winding: 660vAC RMS!
HT mA: 400

Secondary winding resistance is unknown..

Sure I could email Sowter, but I'd much rather ask you how to calculate the proper operating (after load) voltage for myself..

See, Sowter's closest trafo is 660vAC RMS at 400mA, if I go and buy one for two frontends, I'll be drawing 200mA or thereabouts which will considerably change the voltage.

I /really/ want to use Sowter trafos they're the closest thing I've got to a 900v B+

Anyway, heres the rest:

CLCRC

Diodes: Open to suggestions. PIV needs to be min of 2000v
Plenty of tubes categorise for this but I'll also need to calculate for voltage drop with those!.

C1: 33.5mF (16mF 660vAC + 17.5mF 880vAC = 33.5mF)
L1: Plitron PAT-4173 10H, 50 ohm (PAT-4173 pdf)
C2: 33.5mF (16mF 660vAC + 17.5mF 880vAC = 33.5mF)
R1: 4x 20 ohm 3w (film or carbon?) Resistors in parallel
C3: 33.5mF (16mF 660vAC + 17.5mF 880vAC = 33.5mF)



An isolated bias and (both, bias and filament) regulated CLC 10v filament supply.

Cheers.

 

[Sch3mat1c]

Er, your R isn't going to do squat. And where did you find 16,000uF caps rated for that voltage?

So um, what was the issue? Transformers are rated either for AC, or DC output with a certain type of rectifier (usually cap input). In worst case, your transformer will be at full ratings with approx. 250mA DC. It will get hotter if you use excessive capacitance, but >220uF is unlikely to be seen here. And you're going to use what, 200mA (= 160W?) per channel? If so, then hopefully it's 400mA *DC*. If 400mA AC, you'll need two of these things...

660V will run from approx. 933V open circuit to maybe 850V full load, depending on resistance. I would expect 800-900V with tube rectification, 900V typical for silicon. You need >1800 PIV rating, 1A will be fine (three 1N4007's in series per leg).

Tim

 

[Layberinthius]

Sorry I was using an old definition of "uF" which was commonly displayed as mF in books/mags/etc... I've been reading too many 1960's UK magazines by the looks of it

Nethertheless,

mF = millifarad, uF = microfarad..

Report me for being indecently charged...=P

Microfarad is what we want =)

I didn't think that in cap loaded trafos the amps were different in converting from AC to DC (When talking about post-rectifier), I thought amperage was defined as DC after any capacitor and only when dealing with choke loaded supplies was the amperage value changed.

OH I see now, this is a centre tapped transformer, not bridge rectification.

I believe you are incorrect.

This is a CLCRC supply, cap, choke, cap, resistor, cap.

No It's 100mA per channel, not 200mA.

 

[Brett]

Calculate it? Bah humbug. Use PSUD. Accurate within a couple of percent and much faster than doing it longhand.

I've used the 813 a lot (my fave tube), but not SE much. It really likes >700V and around 100mA. DC filaments are a painsaver, especially current sourced. If you haven't decided on a driver, try an EL84/D3a/C3m/12HG7 in pentode, with a low anode load. Thorsten's page has a cct called Legacy that uses something similar to what I ended up with, and Pete Millett's has lots of comparisons and other specs on pentode and triode drivers.

If you haven't already seen it, take a squizz at Stephen's 813 journey and Pete Millett's page also has an SE813.

I just picked up a 2600Vct transformer so I'm going to try a PP pentode (with plate to grid feedback) 813 bass guitar amp.

 

[Sch3mat1c]

Sorry I was using an old definition of "uF" which was commonly displayed as mF in books/mags/etc... I've been reading too many 1960's UK magazines by the looks of it

I didn't think that in cap loaded trafos the amps were different in converting from AC to DC (When talking about post-rectifier), I thought amperage was defined as DC after any capacitor and only when dealing with choke loaded supplies was the amperage value changed.

Well, if the transformer is used for AC, it will be rated in terms of AC output. 120:240V transformer for instance, used for AC. If rated 1A output (240VA), it goes down to around 600mA for DC. But if the transformer is rated for DC output anyway, then yeah you can safely draw rated current.

OH I see now, this is a centre tapped transformer, not bridge rectification.

I believe you are incorrect.

Nope, I'm never wrong If it's rated for FWCT, then yeah, you can only draw half when FWB'd.

This is a CLCRC supply, cap, choke, cap, resistor, cap.

Not with five ohms between those caps. Or was that a typo?

No It's 100mA per channel, not 200mA.

Ok, then it sounds like you can do it with just that transformer.

Tim

P.S. Bit of info: CT'd transformers are usually rated "VCT" instead of "VAC", which implies no CT on that winding. So 25VCT = 12.5 - 0 - 12.5V = 25VAC if not CT'd.

 

[Brett]

Further thoughts after morning cuppa....

I'm considering a 200mA 900v un-regulated (but well filtered)
power supply for a combined SE 813 Class-A without a seperated power supply chassis, using:

Sowter M045 Centre Tapped.

HT winding: 660vAC RMS!
HT mA: 400

Secondary winding resistance is unknown..

Sure I could email Sowter, but I'd much rather ask you how to calculate the proper operating (after load) voltage for myself..

But you need to know the secondary resistance to calculate the voltages close.

I /really/ want to use Sowter trafos they're the closest thing I've got to a 900v B+

At 145 quid and postage for 7kg from the UK, I'd suggest emailling Bob at Lancroft and getting a quote for one of his made in Sydney. His OPTs are great and I'd recommend them as well as Lundahls, which can be purchased in Oz.

Hammond makes excellent filter chokes which are cheapish and available locally, as do Lundahl (excellent!!) and Lancroft.

Diodes: Open to suggestions. PIV needs to be min of 2000v
Plenty of tubes categorise for this but I'll also need to calculate for voltage drop with those!.

6CJ3 and 6D22S in that order.

 

[Geek]

Nope, I'm never wrong

I always get an "A" when Tim does my homework

 

[Sch3mat1c]

Re: Re: 900v B+ Silly time.

But you need to know the secondary resistance to calculate the voltages close.

And line voltage, and leakage inductance. And diode performance with chosen cap. 10% is good enough for tube purposes, and a simple rule of thumb like 1.3 times works.

IMHO, I hate PSUD anyway. I like doing everything myself. I can calculate it manually, not to mention manipulate it in my mind and on paper faster than digging out the program and so on. Plus I don't break the keyboard if I get bogus results and electocute myself....(and we all know people have had bogus results from PSUD; a product of the user's programming, not the software's ).

6CJ3 and 6D22S in that order.

1N4007, 1N4007 and 1N4007, in that order

Tim

 

[jrdmedford]

Yes. I am considering taken the plunge to high voltage myself.

You seem to be letting your product preferences drive the design.
I'd not do this with a 900V supply. I think a choke input supply would be
best in this situation. Why? The noise induced by the ripple currents
charging the input cap will be very high with 660-0-660 transformer.
These same ripple currents might drive the transformer into saturation
and cause it to buzz, or get warmer than it should..

To do choke input, you'd use a 1000-0-1000 (2000 VCT) transformer.

The downside to choke input is (1) It must draw a MINIMUM load current
at all times otherwise the voltage rises to 1500V. (2) You need to tune
a snubber network to quench switching spikes. (3) Layout and parts
selection is off utmost importance for safety (both fire and electrocution).

Hammond has a bargain-basement 2000VCT tranny for USD$64.00.
CCS service is 150mA, so you'd have to buy two and do a dual mono
supply. They also have various HV chokes.

This sounds more complicated, but in the long haul, its the best choice
for preserving the integrity of the sound reproduction.

Perhaps Sowter can do a custom HV transformer.

-- Jim

 

[rcavictim]

Hammond has a bargain-basement 2000VCT tranny for USD$64.00.
CCS service is 150mA,

-- Jim

Hammond has become so expensive that I figure a doorbell xfmer would be all you could buy for $64.00. This plate xfmer sounds interesting. What is their part number please? It sounds perfect for large single triode monoblock applications.

BTW, I'm with you on recommending choke input filter. If you can switch the plate xfmer on a delay timer you can solve the voltage rise without load issue without resorting to a power wasting, heat generating power resistor minimum supply load bleeder, although a bleeder is still mandatory for safety reasons.

 

[Geek]

Look in some older electronics stores for some dicontinued Hammond stuff. I walked away with a transformer that had a retail price in 1988 of $108, for only $20 at SMI Electronics, Langley, BC

 

[rcavictim]

Look in some older electronics stores for some dicontinued Hammond stuff. I walked away with a transformer that had a retail price in 1988 of $108, for only $20 at SMI Electronics, Langley, BC

While I always follow the surplus stores/hamfest route to one-of good deals on parts suitable for supplying my own needs I am also looking for possible sources of new parts suitable for possible production runs since I am entertaining the notion of sharing what I seem to have a gift at doing.

 

[Geek]

.....since I am entertaining the notion of sharing what I seem to have a gift at doing.

Please email me - we may be able to help each other

12ax7 <at> telus <dot> net

 

[jrdmedford]

Hammond 728 is the model number

www.angela.com sells them. I correct myself. Not $64.00 but rather
$55.00. Use the savings to buy yourself a nice CT filament
transformer to go with it.

If you look at the specs on the Hammond site you will see that the
HV leads feed through bushings on the endbells. These are aching for
you to attach ceramic plate caps for 866As or your favourite high
vacuum rectifier.

-- Jim

 

[Layberinthius]

Er, your R isn't going to do squat. And where did you find 16,000uF caps rated for that voltage?

660V will run from approx. 933V open circuit to maybe 850V full load, depending on resistance. I would expect 800-900V with tube rectification, 900V typical for silicon. You need >1800 PIV rating, 1A will be fine (three 1N4007's in series per leg).

Tim

It was a typo, it is supposed to be 50 ohm.

According to the Sowter website they specify all of their mains transformers as /full-load/ RMS AC.

So a 660-0-660 at 400mA would be after full wave = 660vDC ?????

Confusing!

Calculate it? Bah humbug. Use PSUD. Accurate within a couple of percent and much faster than doing it longhand.

I've used the 813 a lot (my fave tube), but not SE much. It really likes >700V and around 100mA. DC filaments are a painsaver, especially current sourced.

If you haven't already seen it, take a squizz at Stephen's 813 journey and Pete Millett's page also has an SE813.

I just picked up a 2600Vct transformer so I'm going to try a PP pentode (with plate to grid feedback) 813 bass guitar amp.

Good luck!, oh and I've decided to follow this schematic with the 6SN7, but of course I'm not considering that as a final tube, if I'm impatient to put it in a chassis I'll make up a faceplate style bracket so I can swap out the octal/9pin/7pin sockets easier than soldering them...

Luckily a lower voltage is easier to construct

Well, if the transformer is used for AC, it will be rated in terms of AC output. 120:240V transformer for instance, used for AC. If rated 1A output (240VA), it goes down to around 600mA for DC. But if the transformer is rated for DC output anyway, then yeah you can safely draw rated current.

Nope, I'm never wrong If it's rated for FWCT, then yeah, you can only draw half when FWB'd.

Ok, then it sounds like you can do it with just that transformer.

Tim

P.S. Bit of info: CT'd transformers are usually rated "VCT" instead of "VAC", which implies no CT on that winding. So 25VCT = 12.5 - 0 - 12.5V = 25VAC if not CT'd.

Yep I should be able to no probs, I'm worried though at the IFRM being too high, but I'll keep an eye on it in PSUD...

From what PSUD can tell me a 660-0-660 trafo isn't enough to push it up to /around/ 900vDC, it falls short of 200v, so I re-calculated for 750vAC RMS and hit the spot...

Now heres the problem with Sowter, seeing that they claim that their volt ratings are full-load and I'm only going to be using what? 2/3rds of 500mA, the final voltage will be much higher?

The closest Sowter trafo is 750vCT at 500mA...M056

Further thoughts after morning cuppa....

But you need to know the secondary resistance to calculate the voltages close.
At 145 quid and postage for 7kg from the UK, I'd suggest emailling Bob at Lancroft and getting a quote for one of his made in Sydney. His OPTs are great and I'd recommend them as well as Lundahls, which can be purchased in Oz.

Hammond makes excellent filter chokes which are cheapish and available locally, as do Lundahl (excellent!!) and Lancroft.

6CJ3 and 6D22S in that order.

1#, Sure do, I'll have to get onto Sowter.
2#, Thanks I'll shoot off a message after I'm done...
3#, My original idea was to use Hammond chokes, the 193Q has a max oper of 1000vDC and 500mA DC max current rating, so it fits just perfectly, as long as I don't go over 1kV, which is just 100v away
4#, 6CJ3 or 6D22S last, it is...

I want to know what you guy's think..

Cheers!

 

[rcavictim]

Jim,

That is a great price. These have to be NOS as Hammond discontinued making such xfmers many moons ago.

Yes, with the leads designed to go directly to a pair of 866A mercury rectifiers these would really give your amp the retro look!

I recently used the Hammond 720 with full wave SS diode bridge in my re-engineering of my Ming-Da 805 SET stereo amp.

 

[Layberinthius]

Jim,

That is a great price. These have to be NOS as Hammond discontinued making such xfmers many moons ago.

Yes, with the leads designed to go directly to a pair of 866A mercury rectifiers these would really give your amp the retro look!

I recently used the Hammond 720 with full wave SS diode bridge in my re-engineering of my Ming-Da 805 SET stereo amp.

I kind of tend to agree with fdegrove on MV diodes, they're not only noisy on the PSU, I wouldn't be surprised they also let out RF noise, my region is an RFI emission free region as it is, I didn't hoist up many an antenna for nothing

It's a ******* great idea, the combination of these tubes would make the amp look like it went on SpaceShipOne, but but but...

If I was going to use a pair or quad (paralleled) of 866A's, I'd have to stick them behind a faraday cage, like a many HV deflector circuits in old vacuum tube tv sets...and ground like crazy!

 

[rcavictim]

Benjamin,

It is true that Hg rectifiers are RF noise generators. You are supposed to use snubbing network filters in the anode leads to help suppress the noise. A shielded winding plate xfmer would help here too although since it is an isolated xfmer there are options to filter the primary before it gets to the rest of the amp. I haven't used Hg rectifiers yet in an amp but was considering it with my 813 project. My plan was to place the rectifiers physically away from and behind the chassis top iron from the audio chain tubes, and also to use shielded coax to run to the anode caps. The choke filters for the anodes would be just under the chassis.
The RFI Faraday cage is the best idea but will likely detract from the desired artisitic look of the finished amplifier, if you are of the same school of thought as I am, that a domestic audiophile tube amp can also stand out as a piece of decorative art.

 

[Knarf]

Getting in a little late for this discussion, but...

Well, if the transformer is used for AC, it will be rated in terms of AC output. 120:240V transformer for instance, used for AC. If rated 1A output (240VA), it goes down to around 600mA for DC. But if the transformer is rated for DC output anyway, then yeah you can safely draw rated current.

Nope, I'm never wrong If it's rated for FWCT, then yeah, you can only draw half when FWB'd.

...this was too good an opportunity to pass by.

Tim, you are wrong.

Ladies and gentlemen, please turn your attention to the middle of page 1175, chapter 30 in Radiotron Designer's Handbook, 4th edition, 1956 by RCA.

There is a graph, showing the ratio of RMS rectifier current to average current (ie. rectified DC), given different combination of capacitor, load resistor, ripple frequency and transformer/rectifier loss resistance. The ultimate limit for this graph is a value if 1.584, the inverse of which is 0.63, the value Tim mentions.

However this value it is hardly possible to reach in any practical design, unless you want to torture the choke with high ripple voltages, ie. make the capacitor so small, that the design approaches a choke input filter.

For all reasonable values of 2 * Pi * Fripple * C * Rload, combined with sensible transformer and diode loss resistances (ie. rather low to ensure a high output voltage), the ratio is between 2.6 and 5.6 in the ultimate top limit (very large capacitor-to-Rload ratio, very low transformer and rectifier losses).

So for a load current of 100mA, the transformer should be 'overrated' between 2.6 and 5.6 times in its AC current rating in order for it not to be run outside specs, meaning 260mA and up to 560mA. The reason, of course, is that the peak current drawn when charging the capacitor is much, much higher than the average load current, so even though the pulses are comparatively shorter as compared to running DC through the trafo loss 'resistor', the losses are proportional with I^2 and they catch up with us.

Many DIY electronics projects badly underrate the transformer for this reason, and, as we are talking class A here, we really have no slack to give when specifying the transformer. Keep all of this in mind the next time you hear someone complaining about hot HT iron.

PSU designer wannabe's could do worse than reading all of chapter 30 in Radiotron. No computers needed, just lots of nice, friendly graphs. A pocket calculator may come in handy though, when you want to relate the theory to real world components values...

Frank.

 

[Layberinthius]

Benjamin,

It is true that Hg rectifiers are RF noise generators. You are supposed to use snubbing network filters in the anode leads to help suppress the noise. A shielded winding plate xfmer would help here too although since it is an isolated xfmer there are options to filter the primary before it gets to the rest of the amp. I haven't used Hg rectifiers yet in an amp but was considering it with my 813 project. My plan was to place the rectifiers physically away from and behind the chassis top iron from the audio chain tubes, and also to use shielded coax to run to the anode caps. The choke filters for the anodes would be just under the chassis.
The RFI Faraday cage is the best idea but will likely detract from the desired artisitic look of the finished amplifier, if you are of the same school of thought as I am, that a domestic audiophile tube amp can also stand out as a piece of decorative art.

Well seeing we're gonna go crazy here, why not the 575A?

It's just as big as an 813, you need two, so why not go crazy?

Imho the 6SN7, fireflies, mosquitos will be a problem, not to mention radiation