High voltage Class D

[nonoise]

The idea calls for an amp that can deliver 600V P-P at say 200mA
The problem with most driver chips is that their driver stage seems difficult to mate with high voltage output devices.
Hopefully I’ am wrong, Comment would be most welcome

Some additional info
Analog input (line level)
THD+N 0.1% good enough

[phase_accurate]

There are drivers available form IRF that go up to 1kV.

Keep in mind that you'd have to struggle with parasitics more than if you were to build an amp with lower supply voltage.

OTOH the fireworks probably get more interesting !!!

regards

Charles

[Ouroboros]

I had (and still have) an interest in designing a UCD-style Class-D amplifier capable of driving 100V-line loudspeaker systems without requiring a step-up transformer on the output. This requires a amplifier that will swing in excess of 300V pk-pk at the output. Fortunately it can be a bridge amplifier so each half of the bridge is swinging less than 200V pk-pk even allowing for sufficient output voltage headroom.

The load impedance presented by the 100V line will be 50 Ohm or greater (depending on the power-tapping at the speakers) and so the output inductor value in the amplifiers is around 120uH rather than the 20uH normally expected, but even so, there are serious issues in providing fast-acting current limiting in the amp making it short-circuit proof.

[ChocoHolic]

At 600V you should definitely not try to run the body diodes of the MosFet in hard switching conditions.
This is usually fatal.

You should spend two ultra fast diodes for each MosFet. One in series to the MosFet, making sure that the internal body diode does not carry any current.
A second across the series connection of the MosFet+1st_diode.

A suitable driver would be IR2113.

But take care. 600V tend to kill people who do not take care, or who are just slightly unconcentrated or tired ... !
Usually in professional labs the engineers are not allowed to work on such hardware if they are alone.
Min 2 people in the lab and easy to reach emergency turn off buttons. Use an isolation transformer. Do not touch at all, when in operation. Use a fast bleeding resistor for the rail caps. Wait 5xR_bleedxC_rail after turning off before touching. Use one hand only.
If you follow these rules you have a chance to survive your experiments. But even then you should settle a proper life insurance first to cover the needs of your family.

[BWRX]

Originally posted by ChocoHolic
But take care. 600V tend to kill people who do not take care, or who are just slightly unconcentrated or tired ... !
Usually in professional labs the engineers are not allowed to work on such hardware if they are alone.
Min 2 people in the lab and easy to reach emergency turn off buttons. Use an isolation transformer. Do not touch at all, when in operation. Use a fast bleeding resistor for the rail caps. Wait 5xR_bleedxC_rail after turning off before touching. Use one hand only.
If you follow these rules you have a chance to survive your experiments. But even then you should settle a proper life insurance first to cover the needs of your family.

That is a very good warning about working with higher voltages. Safety should definitely be considered first. If you don't have the proper instruments, tools, and workspace please do not attempt working with such dangerous voltages. We want our members to enjoy discussing, building, and listening to audio equipment. And to do that you must have a pulse.

[fredos]

easy way...6N137 opto with TC44XX serie driver....With good high voltage bootstrap diode...Et voila!

Fredos

[darkfenriz]

Markus
diodes will switch at around 600mV, while load current should be limited to 200mA as design requirement, I assume very little conduction spikes are also mandatory!

Isn't it like any mosfet with Rds-on below 3 ohms prevents from switching of body diodes?
Or more realistically- can a body diode switch on when Rds-on is like 0.2ohm with 200mA current?

[ChocoHolic]

Hi Dark,
...not really...
Imagine the current path during dead time.
Situation:
Lower switch ON & output current running from the halbridge output into the inductor.
(Typical scenario during large pos half wave of music signal and lower switch ON)

Then switch OFF the lower MosFet. The current will immediately commutate to the body diode of the MosFet during dead time.
After dead time you turn ON the upper MosFet, which now is delivering by brute force the reverse recovery current peak for the body diode of the lower MosFet.
It will depend on the choice of devices and the designed turn on speed if you just get heat or smoke.... with high chance for smoke.

And regarding smoke: Smoke in power electronics can also mean splattering components. Eye protection is sometimes a good idea.

[nonoise]

Thank you all for the replies
The next problem would be to find a suitable analog input modulator
All Ti device seem to be PCM only, perhaps Trypath but I’ am not sure
If one could modify it using the IR driver
Comments ?

[ekaerin]

Hi there,
Switching high voltages at high speed involves losses due to charging/recharging
capacitances etc. For a kW class converter this is no problem because the output power is still high compared to switchlosses and
the switching frequency probably a lot lower.

Maybe an option to investigate is the use of a step up transformer to get the desired output voltage.

regards / Mattias