UCD180 questions

[rosbacke]

Hello out there.

I plan to buy two Ucd180 modules. I have a powersupply which delivers +-60V DC / 5A unregulated with 2x10mF on each rail which I plan to use.

My plan is to build two simple linear regulators using a power transistor, zener diode, resistor and a capacitor to lower the voltage to around +-48V before supplying the Ucd180. The circuit will be very simple and intended only for protection of the module.

Could this cause any problems? Any special things to consider when building this? Should this circuit be placed before, in between or after the capacitors?

Grateful for any feedback.

--- Mikael Rosbacke

[classd4sure]

Hi,

A problem I see is the loss of efficiency you'll have by using a linear regulator in a class d circuit.

Are you doing this to make use of the transformer you already have, or to run the modulator up to it's limite safely?

I doubt it would be necessary for protection purposes if you left a bit of headroom on the rails as previously mentioned.

[rosbacke]

classd4sure, thanks for the answer.

The reason is that I have an unused power supply already assembled and want to avoid buying a new one. I'd figured I will still get around 240W of power to the two modules. It should still be plenty for my flat to turn into a disco...
I know I will lose energy in the linear regulators but I can live with that. But I'm uncertain if the regulator will have some degrading effect on the sound or if the module draw power in such a way that my regulators can affect their performance. Hence the question.

--- Mikael R

[UrSv]

I'd definitely wait for the UcD400 which runs at 60 V and abandon the PSU idea. Cost is +EUR 40 each meaning +EUR 80 in total. Power is higher and less work and simpler to get running alright.

[classd4sure]

Hi,

I like your approach to the situation, making what you have work.

Soundwise my guess is if your regulator is a winner there's a good chance it sounds even better due to the decreased ripple >psrr

Make sure your regulator is a winner.

You mentioned using one resistor and one cap, is that to bias and filter the zener current? I think that's two resistors and you need another power resistor to limite the current through your pass transistor.

You could look into using an active current source to feed the zener with, won't need to filter it then.

I'm certain you can find all kinds of variations on the circuit to look at here that could give you some ideas.

You shouldn't have to make anything too extravagant but with the few parts you mentioned I feel it's a bit too simple, at least add the power resistor.

You might want higher current capability in the end but it will get you started.

[classd4sure]

Originally posted by UrSv
I'd definitely wait for the UcD400 which runs at 60 V and abandon the PSU idea. Cost is +EUR 40 each meaning +EUR 80 in total. Power is higher and less work and simpler to get running alright.

That's an option too but .... I wonder what kind of power his speakers can handle. That would be a small supply for it, you don't want to skimp on the current capability, it's a small supply for a 180watt system. No point paying for the extra power if you can't put it to use, that's just throwing money at it for nothing, and a regulated supply could improve things if it's done right, less than 1/4 the cost of two 400's.

In the end I think his money would best be spent on a proper transformer of the correct voltage rating and muuuuuuch higher current capability. People seem to like the 1000VA ones for instance, but trying it with linear regulators first would be alot cheaper than that and once he gets to hear it he'll be more apt to want to fork over more money for a decent transformer.

I don't think 5amps will cut it at all. These modules should be worth building a good supply for, but if ya can't swing it right now go for the regulator to make the lights glow and I hope you get to hear something from it too, I think it will work, but it won't come to life until you get a real supply for it.

[rosbacke]

Thanks for the replies.

UrSv, It is a tempting idea. But i don't think it will solve my problem. The absolute maximum supply rating of the 400 module is +-60V, i.e. there is no margin at all. Even I would like maybe 5% safety margin.
But my main problem is that two 400W modules could potentially draw 800W of power. This could really take my 2x160VA transformers over the top.

classd4sure, A good transformer is the right solution in the long run. But currently there are other parts of my system in need of an upgrade before I can really appreciate the full potential of this module.
The speakers are fairly OK (Two way using Vifa PL18 and DX25 drivers) but I don't think they can handle more than 100-150W of power. The source is my main problem and the next DIY project. It will be some kind of DAC so I have a good signal to start with.

I havn't done the details yet of the regulators. I made a point of keeping them simple but given the voltage 48V and currents of 5A I will probably need to use some additional small signal transistors and capacitors unless the zener is going to handle a lot of power.

Anyway, I'll have to view the regulators as a temporary solution until I can justify a new transformer.

--- Mikael R

[UrSv]

I think it would solve your problem completely. ..

Sure your modules could potentially draw even more than 800 W but that's when YOU draw that from them. With music and keeping reasonably well away from frying your speakers you will get nowhere near. Besides, if I'm not mistaking the recommendation is 60 V but they will take 63 V easily.

I know it's controversial and there are many views and beliefs on the subject of transformers for Class D amps but IMHO I think it's clear that if you don't play them flat out then a proportionally smaller transformer will *suffice*. Lars at LC Audio (ZAPpulse et al) tested to run their module with something like 55 VA for a 200 W module with very good results if my memory serves me correctly. Sure it was an analog PWM supply but still. I'm sure that will mean some will tell you he hasn't got a clue but after all he makes a living doing these things. Also, the fact that a transformer is rated at 320 VA only means that's the power when internal temp rise is according to spec. It will probably happily let you have 800 VA when asked for shorter peaks.

Just my two öre.

[classd4sure]

Hi,

Sounds like a working plan you have there to me.

Regarding your source, I really like new audigy's... it isn't very diy though but ... hard to beat the quality and options for the price of them, thx certified aside they sound great. It's enough to overlook the low quality outputs used on them. I wonder how others feel about that. Kind of off topic though. Sorry

[Bruno Putzeys]

Originally posted by JohnW
D/S over-voltage even for nS’s (Ringing) will break down a FET.

This is indeed the case if the spike is delivered by a low-impedance source, like a test generator. If this source is capable of delivering large currents, the avalanche mechanism is indeed destructive.

This is not the case with ordinary overshoot/ringing. The energy "stuck" in the ringing corresponds to what can be stored in the stray inductances (pins, bond wires) at the current that's flowing (0.5*L*I^2). When this energy is released across an open FET it'll go in avalanche, absorb the impulse like a zener and come out of avalanche when the stored energy has petered out and the voltage is returned to below-breakdown. The single-pulse maximum one can dump into a FET is actually given in the data sheet.
This is not to be taken for granted btw, MOSFET designers take special care to allow absorption of impulse energies without turning on the parasitic bjt. This is what they call "ruggedness". Current day mosfets are orders of magnitude more "rugged" than what is required to absorb the kind of spikes found in a class D amp.

If a MOSFET breaks down gate oxide before going into avalanche this is called a case of bad fet design...
If you have experienced sudden breakdowns near maximum operating voltage (and current) this was more likely caused by parasitic bjt turn-on due to excessive diode recovery currents. Ringing on the actual power rail (which encompasses electrolytic's parasitic inductance and the local ceramic cap) is another thing the FET may not be too happy about.

Cheers,

Bruno