Megatron Electrostatic Headphone Amplifier

[dennis h]

Questions

 

What is the overall gain of the Megatron ?

 

What is the voltage swing ?

 

thanks in advance,

dennis h

[Earspeakers]

I came to this board late ... are there any more boards for sale, or is anybody planning a new run?

[JoaMat]

/

Edited November 11, 2015 by JoaMat

[kevin gilmore]

seen a bunch of the ones on the left, every one tests as a darlington of about 30 volts

 

if you got them from dalbani, or people associated with them, definitely fakes

Edited February 12, 2015 by kevin gilmore

[JoaMat]

/

Edited November 11, 2015 by JoaMat

[kevin gilmore]

fqpf8n80c is an insulated part and goes directly on the heatsink

and with a metal screw.

 

power supply runs a lot hotter if the unregulated voltage is too

high due to the transformer voltage.

[JoaMat]

and what about the normal temperature of your heat sinks.

 

I don't know - I used a PSU with external heat sink. Maybe Eggil or some other Megatronist  can answer that.

[spritzer]

It's not recommended to run the Megatron with a KGSSHV psu and if you want to do it, even the correct 2.5" sinks just don't cut it.  Way too much current. 

[mypasswordis]

Speaking of power supplies, I tried firing up the BATE power supply (should be same as KGSSHV power supply but without the error amplifier and with an additional fixed bias line [edit: instead of the front end supplies]) for the first time. Fried a 2k resistor because the trace went to the cathode of the 10m90s instead of the gate, so I cut the trace and re-routed that and now the +400V rail and bias works fine but I'm measuring that at some point GND is tied to -430V so the negative supplies don't work. Any of you know where exactly this point is? I looked through all the other traces after the 2k resistor mishap and everything else seems to be right.

Edited February 12, 2015 by mypasswordis

[spritzer]

I looked at the original files and yup, found the issue.  We updated the layout software around that time and they changed one small detail, the relief for the pins followed when doing copy/paste which is what happened here:

 

 

See the yellow rings, they are attached to the plane but really shouldn't be.  I'd just remove the caps and drill into the top, then insulate. 

[mypasswordis]

Thanks! That was fast. Was looking into buying a drill press anyway.

[eggil]

This is the kind of heatsink you want. Big enough to dissipate heat. An external one worked pretty good for me with temperatures not higher than 50 F or so.

Edited February 13, 2015 by eggil

[nopants]

I think there's a cuboid case out there in china where only one of the sides is a heatsink, this is probably ideal.

[eggil]

I could not find one the right size for what you mentioned and ended up using this
http://www.ebay.com/itm/BZ4309-Full-Aluminum-Enclosure-Both-sides-heatsink-power-amplifier-box-chassis-/141252805356?hash=item20e352c2ec

Edited February 13, 2015 by eggil

[mypasswordis]

Might as well replace all the pass FETs while you exchange the heatsinks for the right ones, remove the thermal washers, use metal screws and nuts, and put some thermal paste on. Also 420VAC secondaries is way too high for 450VDC output.

[JoaMat]

/

Edited November 11, 2015 by JoaMat

[mypasswordis]

The melting point of nylon 6,6 is about 250 degrees C or 500 degrees F, which was apparently reached outside of the transistor case. The max junction (internal) temperature is usually 150 C.

[kevin gilmore]

suggest you remove the current limit transistors, and absolutely the correct heatsinks

which are about 3 times the dissipation of the ones you used. use thermal grease

on the fqpf8n80c, mount directly to heatsink with metal screws.

[mypasswordis]

I wouldn't remove the 2n3904s unless they're fried. You do need to change the current sense resistors to 10 ohms. Get a new transformer with lower secondaries, replace all fqp8n80c with new ones, and you should off-board them to a large heatsink like eggil did. Don't forget the thermal paste.

 

Edit: Oops, typed too slowly. KG has spoken. Do what he says

Edited February 15, 2015 by mypasswordis

[nopants]

With the 3904s the amplifier board is current starved and the rails start to sag, to what value I can't remember

[RiStaR]

Thank everyone.

 

i will order some 10Ohm Resistors and large sinks.

 

But during the Chinese New Year, almost all the shop will be closed.  

 

I have to delay the project.

 

I'd really recommend going offboard sinks... with the right size heatsinks for the KGSSHV board, the temps are still a toasty 90C after 15 mins. This had the side effect of "killing" my LT1021s (it always made the voltage to +430V over time until I offboarded the cooling) when cased up.

 

It's been close to a year and the whole setup still runs well.

 

EDIT: just realized I listed the wrong IC and double quoted heh

Edited March 19, 2015 by RiStaR

[RiStaR]

Is there a reason why you're still using the ceramic insulators? The fqpf8n80c are insulated... so they actually reduce thermal conductivity. With regard to your balance issue are your tubes matched? Also check your attenuator (I had an issue once where the relays weren't working properly on my attenuator).

 

Also, how do you like the sound?  

[spritzer]

The 21K is to drop the +450V down to +300V to power the front end.  There is no attenuation anywhere so either wrong parts somewhere or the tubes are bad. 

[RiStaR]

By attenuator, I meant your volume control (pot or relay-based)... but it's likely your tubes especially if they aren't matched. By lazy and temporary, are you saying you didn't use thermal paste between your transistors and your heatsinks or the ceramic insulators? Probably not a good idea there...

[nopants]

from the picture it looks like you are only off boarding one of the fets?