MOSFET LTPI

[Zozobra]

Hi guys, I know there have been a few discussions about using MOSFETs at phase inverters around here but most of them pertain to cathodyne types and the few on the LTPI types all seem to have lost the schematics.
Anyway, I was looking for some stuff on using LND150's to construct a clean channel earlier and stumbled across this design from an amp someone designed called the deadplate to drive an ecc99 power amp.



It looks like a straight sub in and could be pretty good for high gain amps where maybe PI dirt isn't wanted/needed. I guess I could add NFB in the usual way to include depth and presence controls, maybe tweaking the NFB resistor to suit. Is there anything glaringly obviously wrong with the circuit? I'm only just getting to grips with using FETs so I'd appreciate someone casting an experienced eye over it

[Gingertube]

At first glance it looks OK.
D1 and D2 are protection diodes. I would check the LND150 Data Sheet for max Vgs specification. Those diodes may be better placed directly between the mosfet gate and source. 1 zener ONLY cathode to gate, anode to source for each mosfet.

Also look at this thread for Blackstar version of ORDINARY (enhancement mode) MOSFET (not depletion mode mosfet like LND150) Phase Splitter and output tube bias handler.
http://music-electronics-forum.com/t39309/
It has a HEAP of protections wired around it which puts output tubes in to very low idle current when no speaker or guitar is plugged in. You don't need all that although you could retain the control function and just drive it with a "STANDBY" switch instead of all the sensing logic. It is described down the thread a bit.


Cheers,
Ian

[Zozobra]

Thanks, I shall have a read through. They really went to town on the protection circuit!

I did wonder why the design had omitted the usual gate-source protection diodes. I thought they were possibly used to limit the signal swing into the gate to stop the FETs from clipping? However, the LND150 headroom is much lower than 15V from my limited understanding of how to interpret the loadline.

The max Vgs is 20V and from the LND150 fx return circuits I've seen most opt for 12 to 15V back to back Zeners between the gate and source, although the loop used in the jet city amps omits this and I've not heard of anyone complaining that the loop has failed on them on the various forums I go frequent. Still, it's better to add them for the whole 4p they'll cost to add.

[Zozobra]

One quick knuckle headed question about depletion mode MOSFETs.

I understand that the device is normally on if Vgs is greater than some threshold value, which is ~ -1V for the LND150.
For the biasing arrangement in the schematic I posted above using a source resistor what happens for Vgs > 0? Does the device clip? Thus the gate swing head room is limited between 0 and -1V and must be padded down to this level?

[MarkusBass]

Interesting circuit - have you got full schematic of the amp? I think that it will work but it will pass only quite low amplitude of the signal (about 2.5V RMS). That's why it was used in low power amp.
You may need to adjust 47k resistor in first MOSFET (in the same way as it is done in tube circuits). Maybe with some adjustment it will pass higher signal. The gain of the circuit is about 19dB. Let me know if you build it.

Mark

[nickb]

One quick knuckle headed question about depletion mode MOSFETs.

I understand that the device is normally on if Vgs is greater than some threshold value, which is ~ -1V for the LND150.
For the biasing arrangement in the schematic I posted above using a source resistor what happens for Vgs > 0? Does the device clip? Thus the gate swing head room is limited between 0 and -1V and must be padded down to this level?

No. When Vg>Vs it won't conduct like the grid of a tube does. Therefore you won't get the same time varying duty cycle modulation that is associated with a tube LTPI. It will be interesting to see which you prefer.

[Zozobra]

I found it on el34world. Schematic can be found here:
La Révolution Deux: DEADplate - "five to nine" - Practise and recording amp

[nickb]

Did you see the note on them that says "All voltage measurements +/-100%"?

[Zozobra]

Did you see the note on them that says "All voltage measurements +/-100%"?

Best spec the filter caps for 800V then

[Zozobra]

Interesting circuit - have you got full schematic of the amp? I think that it will work but it will pass only quite low amplitude of the signal (about 2.5V RMS). That's why it was used in low power amp.
You may need to adjust 47k resistor in first MOSFET (in the same way as it is done in tube circuits). Maybe with some adjustment it will pass higher signal. The gain of the circuit is about 19dB. Let me know if you build it.

Mark

How did you calculate the gain/output of the circuit? I'm wanting to learn how to design with HV mosfets but I don't know enough about them to know what to look for to learn about them. I've read the LND150 stuff on hawestv.com but I'm still a little confused after trying to draw some load lines myself :/

[Gregg]

I found it on el34world. Schematic can be found here:
La Révolution Deux: DEADplate - "five to nine" - Practise and recording amp

I hope nobody is building that "thing" because it has many things done wrong.

[PeanutNore]

I'm interested to see how this turns out. I just wired up a source biased LND150 cathodyne PI (source-o-dyne?), because I didn't want to worry about how much Vgs swing it can handle and this should have the source voltage following the gate pretty closely so Vgs should never be very large at all. I'm using a 330 ohm and 47k ohm resistor pair in series from the cathode to ground, with the gate referenced to the bottom of the 330 ohm through a 1M - basically the cathode biased cathodyne from Valve Wizard with the FET substituted in. Set up this way it may not even actually need a gate protection diode, but I've included a 12v zener to be safe.