Antek Power Transformers

[Tage]

So I hooked the transformer to a 8.3V ac transformer and got the voltages I expected. Yellow=430 Grey=70 and white=0, they all read a little higher, as they were unloaded. I had wired all these together in parallel, yellow to Yellow, etc... The Heater windings measured in at 6.4 a each, but I'll run them as a center tapped 12.6, center tap connected to a humdinger pot.

So if I wire the yellow and white to a bridge rectifier, I will not be able to use the 70V tap then? If I wire the yellows as a full wave rectifier and ground the whites, I should then be able to use one 70V tap to build a bias supply correct? or am I missing something...

[imaradiostar]

I don't think you'd connect the center tap to a humdinger- you'd connect the center leg of the humdinger to ground (or even somewhere about ground, floating positive) and connect the outer lugs of the humdinger to the filament leads- 12.6vac in this case.

I don't fully understand what they intended with the 70v taps on the Antek PT's. If you're using the PT with a bridge rectifier which is the most efficient use of the windings the 70 volt tap is referenced to ground one moment and B+ the next. If you use the PT with a full wave center tapped arrangement you won't get as much current out of it but you could use the other taps as intended.

What kind of output valves are you using?

For KT88's or 6550's I'd use the full winding (470 volts) and bridge rectify the whole thing for around 650 volts. I'd add a winding for bias (easily done with a toroid) and if needed use an additional tube or a mosfet for a screen voltage regulator with the gate fed by a voltage divider so that screen voltage tracks plate voltage. It should allow maximum power from big bottles.

With 6L6's I'd probably connect the PT using a FWCT arrangement with the 400v taps. It isn't the best use of the antek PT and the voltage is on the high side but it should be OK.

jamie

[Tage]

Yes you are correct on the humdinger wiring.

I will be using 6550's and so I will use a bridge rectifier, as far as bias I think I will make a capacitor coupled supply. I know it's not ideal but I have yet to see one fail in a commercial amp. Is there a reason you would regulate the screen supply? would it not still track the plate voltage if it was simply derived from a vlotage divider?

[defaced]

A voltage divider will be too high of impedance. It will sag alot under load. As I think about it for a second, I think if you feed the MOSFET via a voltage divider, it will bounce around with the B+, but just at a lower voltage. Definitely worth the 3 bucks in parts to try out.

[Tage]

So how do I go about biasing my Mosfet then? I'm pretty new with silicon stuff, I've only done a few pedals and mostly op amp stuff. My FET and BJT knowledge is mostly just switching purposes. You wouldn't happen to have a schematic excerpt that illustrates this?

[imaradiostar]

A voltage divider will be too high of impedance. It will sag alot under load. As I think about it for a second, I think if you feed the MOSFET via a voltage divider, it will bounce around with the B+, but just at a lower voltage. Definitely worth the 3 bucks in parts to try out.

What he said.

As for bias supply- I don't mind a capacitive supply on lower powered amps but i'd hesitate to use it on a 6550 powered amp because of the need for a lower impedance bias supply. It'll work but I'd say it's far from ideal. Merlin talks about there being a practical limit for the cap bias supply in his power supply book.

Jamie

[Tom Phillips]

Tage,
You mentioned a cathode biased 4 el34 type amp back in post #13. Now that you are talking 6550's how much power output are you shooting for? Are you planning 2, 4 or maybe 6 6550s?
Tom

[defaced]

So how do I go about biasing my Mosfet then? I'm pretty new with silicon stuff, I've only done a few pedals and mostly op amp stuff. My FET and BJT knowledge is mostly just switching purposes. You wouldn't happen to have a schematic excerpt that illustrates this?

This is where I would start with the design. Basically, the voltage that appears at the gate ends up also appearing at the source. So, if you feed the MOSFET a voltage, it spits it out the back end, but it can pass appreciable current, so it's similar to a cathode follower, it's a current buffer.

The two resistors are the voltage divider. The zener diodes are to protect the gate. They're probably not needed for this application, but for 5 cents, you can protect your MOSFET. Their voltage rating needs to be less than gate/source breakdown voltage as specificed on the data sheet. The MOSFET is just one I grabbed, I don't know if a IRF840 would work or not. You will want one that can handle the voltage, power, and current needed for what you're doing. It will also need to be properly heat sinked. I am not sure if you'll need the decoupling cap or not. Try it without, if you get hum in the supply, then add it back in. I'd start with a value that's used for elevated heaters, so like in the 22u to 47u range.

[imaradiostar]

I would add a gate stopper resistor between the gate and the junction of R9 and R10, just in case.

I would also tend toward a mosfet with a built in zener so you don't have to be as careful handling it.

These are cheap enough and easy to use: STW9NK90Z STMicroelectronics MOSFET

jamie

[Tage]

Yes I am now using an Antek for a 4 6550 amp, that is why I've gone with a higher voltage transformer, I'm looking to generate 150+ watts output from the quad. I have another thread going on it, UL taps and PI Drive.

I may do the whole running another transformer backwards to generate my bias but I've got limited real estate in the chassis.

[imaradiostar]

A small dual primary transformer used backwards from a filament winding would make a great bias transformer, especially if the output grids are driven by mosfets- you'd get a +-160 volt supply if you do it correctly.

There's nothing wrong with adding windings to a toroid! It's pretty easy to do and it allows you to control the voltages.

jamie

[Tage]

All right I like the looks of that, although I will be using the UL taps for this amp so a screen supply wont be needed, but I will be using this set up in a future project.

So the mosfet is used as a buffer to give steady supplied voltage to the screens, and does so as a low impedance source.

I've been looking all over for my copy of Merlins power supply book but I must have left it somewhere strange...

[Tage]

SO for that low impedance mosfet fed screen supply, can I sue the ballancing resistors that are arcross 2 capacitors as my voltage divider? it should be around 1/2 of the HT wich would work well for my application.

[loudthud]

That would be ok except there will be significant ripple on that voltage. About half the ripple on the main B+.

[Tage]

Right so a second Voltage divider with it's own Filter capacitor @ the connection of the 2 resistors. Got it!

I have decided that I like this screen voltage source better than the UL connection I was going to use initially.

[imaradiostar]

You might find you need more screen voltage than 1/2 of plate voltage to reach full power levels with your plate voltage. One of my recent builds used a pot as the top half of the voltage divider so i can adjust the screen voltage anywhere from 1/2 of B+ to all of B+. There is really no reason why you can't have one regulator for the screens and one for the preamp but for many simple designs it's unnecessary. Your call. It'd just be another mosfet and a handful of minor bits.

The cap at the junction of the divider needn't be very big- you can calculate its charge time to select the right value. I'm pretty sure it's t=2*pi*R*C. Since there is such a large amount of series resistance even a .1uf at the junction of a 220k resistor and a 250k linear pot should filter out all the ripple and be almost totally clean while tracking thing plate voltage with a delay of only tiny fractions of a second. I suppose you could use a much larger cap for a "slow turn on" effect but it wouldn't track B+ nearly as well which I see as a disadvantage.

jamie

[Tage]

Sounds good. thanks for the clarification. So I could drop another voltage divider from the B+, maybe separated by a CR filter, to get my pre-amp B+ through another source follower. I like that idea, I can still separate the triode stages with CR filters right just to isolate one from the other.

[Tage]

I finally found my Power Supply design book, and found this mosfet regulator in the series voltage regulators. I'm gonna implement it in this build, although only one, no second one for the preamp as I think it would be unnecessary.

Thanks for the great suggestion Jamie, I thin I will use this in most future builds, in stead of a choke, it's as effective and a lot cheaper...

[imaradiostar]

No problem, happy to help. I have a similar power supply build coming together and I'm planning to feed the screens around 400 volts which, with minimal filtering, will work well for the preamp too. Mine is a cathodyne PI driven by a pentode "loop return" with a preamp that resembles a dual rect dirty channel and a Dr. Z single pentode clean channel. It's gonna be interesting! I'm also planning to implement a mosfet based loop send circuit. The tube count is low- a 6au6 out front, two 12ax7's for the "dirty" channel, and a 6GH8a on loop return and phase inverter duty. No negative feedback so it's got plenty of gain and it's simple. I'm using up junker PT's (voltage is too high, thus the regulator) and OT's but it if works well I'll build one in a real chassis with proper iron. I built it with the thought that it could accept almost any 8 pin bottle- plate voltage is around 550 volts and output transformer impedance is 6k6 so I'm going to probably start with 6L6's and see how it sounds, then go from there.

jamie

[Tage]

Yeah this is going to be used on a bass amp, gonna run a pretty simelar power stage, but probably with some Negative feedback around it just to tame it a bit, as it is intended to be a crunchy bass amp. The pre amp is rather simple, a DC coupled cathode follower running into a FMV tone stack then a parralel triode with adjustable gain on it's bypass cap and on to a 12au7 LTP so I can run it into heavierloads, 56k grid leaks, and a quad of 6550/kt88's

[Tage]

Hey digging up an old thread, but I wanted to inquire about how I calculate the dissipation of the mosfet in the source follower. if I plan to drop 200V do I simply multiply that be my screen+ preamp current to determine the power being dissipated in the mosfet?

[defaced]

Correct. Others will be able to give you an idea of how much, but I suspect that screen current is going to be dynamic, so you will want to add some margin to your calculation.

[Tage]

Yeah I figured it would fluctuate, so I'll just go with a little below the max rating on the data sheets. I have a feeling I'll need a pretty big heat sink

[defaced]

I'm pretty sure you can also parallel MOSFETs on a single node. Or maybe give every screen and the preamp it's own MOSFET. It'll make the circuit much larger and more parts, but distribute the heat dissipation and might make that aspect easier to deal with. You might be able to get away with using the chassis as the heat sink if you go this route.

[Tage]

Yeah chassis as a heat sink always kinda sketched me out. Just because there would be a lot of voltage only separated by a thermal pad and a screw to ground... I suppose I'll see what kinda heat i'm looking at and maybe parallel them up...

[R.G.]

Yeah chassis as a heat sink always kinda sketched me out. Just because there would be a lot of voltage only separated by a thermal pad and a screw to ground

That's a good reason to be wary. One slip or flake of metal from the tab of the MOSFET to the chassis and you have a B+ short if the chassis is properly grounded, and a B+ hot chassis if it isn't correctly grounded. There are many small aluminum heat sinks made to clip or bolt onto the common MOSFET packages, and these can be set up so they're not tied to chassis if you're careful. Several inches of air is a better insulator than a thermal pad.

[Tage]

Yeah just ordered up some beefy heat sinks that I think will do the trick just fine.

[jpfamps]

I have had a real dick ache ordering a PT from Antek. First there was a problem with the shipping fee which was my fault, then the damn thing got lost in the mail, then they were out of stock, then the stock didn't come in! Some 10 weeks or so later they're shipping me an alternative transformer which is probably a little better suited to what I want to do with it anyway. They do suck at communications, especially by email so phone them if you have any troubles. They do offer a lot of transformer for the cash so if this one does turn up I would use them again in the future as really, a lot of went wrong here was bad luck. Just thought I'd share.

If you are in the EU Tube Town do a good range of toroidal mains transformer for valve amps at a good price.

Many of these have bias windings.

The cores are very conservatively rated, and as I've used these with full wave bridge rectification (with the HT windings in parallel) at idle usually comes in at about 10% over the expected HT, so I take this into consideration when designing an amp around these units.

I've not used these transformers with half-wave rectification, and I would expect the HT to be more in line with the expected values.

If I want anything made to suit my specifications I use Tiger Toroids, in the UK, who are actually again very reasonably priced.

[Tage]

Anyone know of some north american custom toroid suppliers? Antek won't do custom stuff...

[Zozobra]

Finally got around to wiring up my AN-3T325 I ordered. They could have made the colour scheme a little less ambiguous for us folks wiring it for 230VAC use!
It seem to work fine although I've only tried it unloaded so far. Seems to pop the mains fuse on second power up for some reason. I'm only using a 680mA fuse though which might be a bit small for a 300VA PT although should that matter greatly if its entirely unloaded?

[jpfamps]

Finally got around to wiring up my AN-3T325 I ordered. They could have made the colour scheme a little less ambiguous for us folks wiring it for 230VAC use!
It seem to work fine although I've only tried it unloaded so far. Seems to pop the mains fuse on second power up for some reason. I'm only using a 680mA fuse though which might be a bit small for a 300VA PT although should that matter greatly if its entirely unloaded?

Toroidals are notorious for large in rush currents, regardless of the loading.

As far as I understand this is due to residual magnetization causing the transformer to saturate when first energized. This is "corrected" after a couple of mains cycles, but can cause fuses to pop.

You are usually advised to use some form of current limiting with larger units.

[imaradiostar]

Put a small voltage to one of the primaries- say 6.3 volts from some other filament transformer. By connecting the primaries in series you can get the phasing correct- but you probably already knew that.

My Antek TF seems to draw a lot of current upon initial connection of AC and it's only a 100VA model. Is your fuse a slow-blow?

Jamie

[Zozobra]

I'll try that to check. I should really scope it too.

Its a slow blow fuse but its pretty under rated for the task I guess. Its fuses are quite visibly blown as the inside of the glass envelope is black! Unfortunately my big box of bits with my fuses in is at my practice space from when I was voicing a big muff circuit a few weeks ago. I'll try cycling it a few times using my current limiter bulb too to see if that helps things.

Thanks for the tips guys!

[Zozobra]

and its absolutely fine with the lightbulb limiter in place. the bulb doesnt light at any point but i guess it tames the inrush just enough to save the fuse.

[Tage]

well i am a couple hours work away from powering mine up, and it's a 400VA. any suggestions on limiting teqniques?