• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

PSUD modelling of a filament supply

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

A 2A3 filament is about 1 ohm, right (2.5V, 2.5A). If I use the defaults in PSUD, the transformer secondary has a DCR of ~ 30 ohms, the filter capacitor has a DCR of 2 ohms, and everything goes totally haywire, the voltage across the load doesn't go above a few millivolts. This happens with a constant current load and a resistor load. So...

* What are reasonable values for the DCR of a 2-10V filament transformer secondary and filter caps in the 5000-10000uF/10-30V range?

* Modelling the current and voltage regulators - the current regulator will have a voltage drop across the sense resistor. Is there a drop inside the regulator itself too? And how do I model the voltage regulator?

* I'm starting to come to the conclusion that there are no Schottky bridge rectifiers that can handle 2.5A, and I'll have to settle for regular silicon rectifiers. Is that correct? If not, could someone please give me part numbers for Schottky rectifiers that I can use for 2A3 filament supplies? I guess full-wave rectifiers would work too, I'd just need a higher output trasformer.

I finally found part numbers for the common mode chokes (I found one that's 150uH at 2.8A - should be good enough, right?) and the regulators themselves, so I think that's everything I need to build this.

Thanks in advance,

Saurav
 
Hi,

* What are reasonable values for the DCR of a 2-10V filament transformer secondary and filter caps in the 5000-10000uF/10-30V range?
Calculate the secondary resistance from the transformer's regulation figures: Typically <1 ohm
Look at a capacitor datasheet: Typically <50m ohm

* Modelling the current and voltage regulators - the current regulator will have a voltage drop across the sense resistor. Is there a drop inside the regulator itself too? And how do I model the voltage regulator?
Allow 4-5 volts drop across a reguator.
* I'm starting to come to the conclusion that there are no Schottky bridge rectifiers that can handle 2.5A, and I'll have to settle for regular silicon rectifiers. Is that correct? If not, could someone please give me part numbers for Schottky rectifiers that I can use for 2A3 filament supplies? I guess full-wave rectifiers would work too, I'd just need a higher output trasformer.

There are plenty of 2 pin TO220 types. Choose a type with an isolated case. :)Just tried to read the number off some of mine, but the eyesights failing:)

I finally found part numbers for the common mode chokes (I found one that's 150uH at 2.8A - should be good enough, right?) and the regulators themselves, so I think that's everything I need to build this.

That's similar to what I've used in the past. I'll be interested in other's opinions.
IMO there is nothing to beat an electrostatic shield between the primary and secondary of the transformer.

Cheers,
 
A 2-pin TO-220 would be a single diode then, right? So I'd have to use 4 of them to build the bridge. I was hoping I could find a 4-pin packahe where someone else had already done that work for me :)

Thanks for the other responses. I don't think the Hammond filament transformers I've been looking at have shields.
 
Saurav said:
A 2-pin TO-220 would be a single diode then, right? So I'd have to use 4 of them to build the bridge. I was hoping I could find a 4-pin packahe where someone else had already done that work for me :)

Five minutes work to make a dicrete bridge from 4 individual diodes. You usually get better single diodes too.

Thanks for the other responses. I don't think the Hammond filament transformers I've been looking at have shields.

None of the Hammonds I've ever seen have a shield.
 
Alright, do these look usable?

http://www.digikey.com/scripts/us/dksus.dll?Detail?Ref=45684&Row=306252

Technical/Catalog Information: MBR735-ND
Diode/Rectifier Type: Schottky
Voltage-Rated: 35V
Current Rating: 7.5A
Package / Case: TO-220AC

So 8 of those, or something similar. The TO-220AC package has that little hole in it - does that mean that I need to individually heatsink these? I didn't see anything in the datasheet about the current rating being valid when used with a heatsink.

And this would be the regulator, which I'd need 4 of:

http://www.digikey.com/scripts/us/dksus.dll?Detail?Ref=47836&Row=293545

Technical/Catalog Information: LT1085CT-ND
Current: 3A
Voltage: Adjustable
Features: Positive
Package / Case: TO-220

And I should probably heatsink these, right? My Pass Pearl phono stage used this heatsink: http://www.digikey.com/scripts/us/dksus.dll?Detail?Ref=48465&Row=77986, so I guess I'll get something similar.

CMC - saw these part numbers posted on an AA thread: http://www.digikey.com/scripts/us/dksus.dll?Detail?Ref=49982&Row=184893. That's 150uH @ 2.8A, maybe I should pick the other one that's 47uH @ 3.5A.

I know I'm being really pedantic here :)
 
diyAudio Senior Member
Joined 2002
Hi,

The trouble with posting links from Digikey is that this doesn't work...

From the description you put in, it should be O.K and when you have TO220 you can often use the chassis as your heatsink.

Check the datasheet before you decide to do this and use some heat conductive paste.

That should save some $ on the budget.;)

Cheers,;)
 
Ouch, I didn't even check that. My apologies.

Using the chassis will get just a little tricky construction-wise, because it'll be a little harder to build it on a PCB. I could find longish pieces of scrap aluminum and attach that to the diodes and regulators, if I lay them all out in a line on the PCB.

On an unrelated note - I was thinking about advice I got here about increasing the first cap in my PS in order to improve regulation. The reason I didn't want to do that was that it would raise voltages. However, now that I think about it, if I got B+ up to the 400V range, I could try the Loftin-White schematic you showed me, with the 6SL7 SRPP direct coupled to the 2A3. That would get rid of one cap in the signal path.

And speaking of caps - given the fact that my speakers show a pretty significant IMD increase below about 60Hz or so, I think it makes sense to keep an input cap in the amp to high-pass everything at around 90Hz, right? The negative impact of that cap should be less than the improved clarity in the speaker's output. I have a polyester or polystyrene in there now (don't remember what I picked up at the surplus store), I'll probably play with other types too.

I'm also thinking of trying better OPTs than the Hammond 125ESEs I have in there... this seems to be a nice enough amp to make it worth the expense. The One Electron UBT OPTs are on sale for $80 each, and if I go with those, I'll need to decide on whether to get the 3K one or the 5K one.
 
Alright, do these look usable?

Yes, but the cathodes of the diodes are the tab on the package so if you're going to mount them on a common heatsink, they'll need to be isolated. Simiarly the LT1085CT has the output on the tab too so same problem. Electrically isolated and a big heatsink should be fine. Well at least I think they will be. I have the flu and am less than coherent.

3K for the 2A3's.
 
Ah, so that's what you meant when you said "isolated" earlier. How do I electrically isolate the tab from the (common) heatsink but still maintain good thermal contact? Are there plastic (or similar) washers that are designed to do this? Maybe I should just buy lots of cheap heatsinks. Or cut up scrap metal into little pieces.

I also saw Schottkys in the more "normal" diode-looking packages - those wouldn't need heatsinks to perform as spec'd, and would take up less space too. So, is there a specific reason you recommended I get TO220-type packages?

And thanks everyone for your patience, I know these are really basic questions. But that's what happens when you have a little bit of theory under your belt but zero practical experience :)

I have the flu and am less than coherent.

Sorry to hear that. I wish you a speedy recovery.
 
diyAudio Senior Member
Joined 2002
Hi,

However, now that I think about it, if I got B+ up to the 400V range, I could try the Loftin-White schematic you showed me, with the 6SL7 SRPP direct coupled to the 2A3. That would get rid of one cap in the signal path.

I like that Loftin-White quite a lot. If you think you can build one with what you have, why not?

And speaking of caps - given the fact that my speakers show a pretty significant IMD increase below about 60Hz or so, I think it makes sense to keep an input cap in the amp to high-pass everything at around 90Hz, right? The negative impact of that cap should be less than the improved clarity in the speaker's output. I have a polyester or polystyrene in there now (don't remember what I picked up at the surplus store), I'll probably play with other types too.

If you're certain the IMD comes from the speakers....?
Still, limiting the bandwidth to 90Hz and upwards will certainly rob the music of its fundament.
Whatever you decide, using a polyester cap at the input of the amp defeats the purpose of building a Loftin-White in the first place.

If you must use a cap, go for the best you can, you can also tailor the cap at the output of the preamp if you use one.


I'll need to decide on whether to get the 3K one or the 5K one.

Using the best OPTs you can get will never hurt and at 80$/EA they sound like a steal to me.
If its' 2A3s you're using then 3K should be your choice.

Hope this helps,;)

P.S. Brett, take good care of yourself.
 
diyAudio Senior Member
Joined 2002
Hi,

I also saw Schottkys in the more "normal" diode-looking packages - those wouldn't need heatsinks to perform as spec'd, and would take up less space too. So, is there a specific reason you recommended I get TO220-type packages?

Single Schottkys would be my choice.

Did the TO220 come recommended by anyone?
I think the single one are much easier to work with and I hear people saying the sound better as well.

If you must use heatsinks for any reason, I'd recommend using one per application iso sharing a single one.

With the tabbed dual Shottkys only the cathodes that are electrically connected to one another can share a common heatsink.

Beware that these heatsinks will carry dangerous amounts a current and voltages here so no poking fingers, please.:att'n:

Cheers,;)
 
If you're certain the IMD comes from the speakers....?

That's what the manufacturer said, and a review/test in a magazine showed it too. I think my speakers are happier if they're not playing any bass.

Still, limiting the bandwidth to 90Hz and upwards will certainly rob the music of its fundament.

Well, I do have a subwoofer. That gets a full-range signal from my linestage's other pair of outputs, and right now I have some strange filter network in front of it (don't remember what slope/XO points I have in there any more) that seems to blend reasonably well with the mains, within the limits of room modes and so on.

Whatever you decide, using a polyester cap at the input of the amp defeats the purpose of building a Loftin-White in the first place.

That's what I was thinking too. So I need the best cap I can get, and between using it at the input vs. using it between driver and output stages, I'm not sure which one will do less damage. The coupling cap will have much mode DC across it, so maybe that's a reason to prefer having the cap at the input? There's another option here, reduce the value of the coupling cap so I do the filtering there. I was advised against doing that though. I can see how it would muck with things if I had global feedback, but I don't. So, the third option might be to use a smaller driver-output coupling cap.

If you must use a cap, go for the best you can, you can also tailor the cap at the output of the preamp if you use one.

No caps at the output of the linestage. If I wanted to use a cap, which would be the 'best' place to put it - linestage output, amp input, or inter-stage coupling? I'll see how much teflon caps will cost me on eBay... 100K input R means I'll need around 0.022uF for 72Hz. Grid leak is 470K, so if I put the cap there, I can use an even smaller value. If not teflon, what else would you recommend? Polypropelene?
 
Ex-Moderator
Joined 2003
Hot under the collar

Schottky diodes are intended for switched mode power supplies, which don't use bridge rectifiers, that's why you buy them as single packages, or, at best, dual.

There are various heatsink insulating kits available, ranging from traditional mica to various plastic types. You need a low thermal resistance (degrees C per W). If you use mica, you must smear both sides of the mica with a thin coating of silicone heatsink grease to fill any voids that would otherwise insulate. Similarly, thinly smear the device and the area of the heat sink to which the device will be mounted.

Inevitably, insulating kits have thermal resistance and cause the device to be hotter than if it was bonded directly to a heatsink. You might think this is a bad thing, but it is not necessarily the case. If you use individual heatsinks, where will their heat go? If they are inside a chassis, they cannot lose heat, so their thermal resistance rises dramatically. Even a Schottky diode will heat significantly at 2.5A. A proper, external, finned heatsink with devices on insulating kits is a better bet than individual heatsinks, no insulating kits, and poor airflow.
 
diyAudio Senior Member
Joined 2002
Hi,

So I need the best cap I can get, and between using it at the input vs. using it between driver and output stages, I'm not sure which one will do less damage. The coupling cap will have much mode DC across it, so maybe that's a reason to prefer having the cap at the input?

Coupling caps work best when biased with sufficient DC voltage.
Putting a cap at the input of the amp will have it no DC bias on it.
Not so good.

By reducing the value of the cap between the driver stage and output valves you can filter just the same.
Remeber this filter is made out of an RC network so always look at the R value following it when doing the maths...smaller caps quite often sound better than bigger ones.

NFB? Is this a SE amp or a PP?
Applying global FB with SE amps is almost impossible to do and with a PP the global FB loop won't care, it only feeds back what it was being fed from the output.

No caps at the output of the linestage. If I wanted to use a cap, which would be the 'best' place to put it - linestage output, amp input, or inter-stage coupling? I'll see how much teflon caps will cost me on eBay... 100K input R means I'll need around 0.022uF for 72Hz. Grid leak is 470K, so if I put the cap there, I can use an even smaller value. If not teflon, what else would you recommend? Polypropelene?

No caps at the linestage? Is it a passive? If not how is DC blocked?

Where to put it is ultimately up to you.
Anyway you turn it some part of the gear is going to en up with it's bandwidth limited.
Normally, I'd put it at the preamps' otput where you could install a second set of outputs, say one full range, the other with a bass cut-off suiting your needs.
That would do the least damage to rendering your gear "my system only"

Cheers,;)
 
EC8010, I see your point, but given my construction skills (or lack thereof), mounting these on an external heatsink will be a pretty complicated proposition :)

Coupling caps work best when biased with sufficient DC voltage. Putting a cap at the input of the amp will have it no DC bias on it. Not so good.

OK, so my assumption was backwards. That's never happened before :)

The amp is SE, no NFB.

No caps at the linestage? Is it a passive? If not how is DC blocked?

It's a passive autoformer. DC blocking... I'm not sure if an autoformer will provide DC blocking, since it's the same wire that connects input and output (this would not be the case with a transformer). There's a cap at the output of my phono stage. I believe there is one at the output of my CD player too. To be honest, I hadn't thought about DC blocking at all.

That would do the least damaged to rendering your gear "my system only"

That's not really a concern. My stuff is ugly enough that I doubt anyone else will want it :) Seriously, I can back these changes out if/when I sell something. Though, I see the logic in what you're saying. Especially since I already have 2 outputs on my linestage (and only 1 input for now).

FWIW, pictures: http://www.diyaudio.com/forums/showthread.php?s=&threadid=18581
 
diyAudio Senior Member
Joined 2002
Hi,

It's a passive autoformer. DC blocking... I'm not sure if an autoformer will provide DC blocking, since it's the same wire that connects input and output (this would not be the case with a transformer).

An autoxformer would be a xformer with a 1:1 ratio.
In this case its sole purpose would be to block DC.For that to happen you do need separate primary and secundary windings.
So input and output wiring must be separate.

Unless the DC voltage at the input is either: neglegibly small ( which I doubt) or carries only a very small amount of current ( more likely) it won't be working well causing core saturation leading to hysterysis distortion.

So I can only assume some measure was taken to avoid DC at its input a la parafeed or the xformer was gapped for SE service.
This is what bottleheads like us call a WOT.;)

I'll take a jump into the link you provided...I can't possibly remember all that goes on here.

Cheers,;)
 
An autoxformer would be a xformer with a 1:1 ratio.

Not in my linestage - it has multiple taps for the output, even though everything's one long wire - the taps are wires that "come out" of the windings, then "go back in" to continue making more turns. So, that's how it works as a volume control - depending on which tap is selected, I get a different number of turns on the output side (the input signal goes across the full number of turns).

So I can only assume some measure was taken to avoid DC at its input a la parafeed or the xformer was gapped for SE service.

Could be the latter, because there are E-I laminations in there that alternate in their interleaving. There's only one autoformer per channel, no AFAIK there's no parafeeding going on. I could ask Dave Slagle about this.

This is what bottleheads like us call a WOT.

What's a WOT?

I'll take a jump into the link you provided...I can't possible remember all that goes on here.

And I don't expect anyone to either :)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.