Hoffman Amplifiers Tube Amplifier Forum

Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Bieworm on July 03, 2021, 08:52:14 am

Title: Vvr mosfet
Post by: Bieworm on July 03, 2021, 08:52:14 am
Anybody Who can confirm this mosfet is a good replacement for the Nte2377?
For a 15 to 20W cathode biased amp.
Title: Re: Vvr mosfet
Post by: vibrolax on July 04, 2021, 02:03:03 pm
I do not have experience with this particular MOSFET, but I have made a few VVR's for ~20W amps.

According to the full data sheet, whose link I provide below, I looked at the Forward-Bias Safe Operating Area graph (Fig. 12 on page 4).  It showed a DC current limit of about 0.3A at 400V, and 0.2A at 500V.   500V @ 200 mA is 100W.  At lower voltages, of course, the current limit will be higher.  Assuming 25% output efficiency, it appears adequate to me.

The case tab is connected to drain, i.e. the DC input voltage, so the device has to be insulated and heat-sinked properly.

I do not claim to be an expert on the subject of MOSFET VVR's, and I welcome more experienced members correcting any mistake in my analysis.

Full data sheet:
https://4donline.ihs.com/images/VipMasterIC/IC/IXYS/IXYS-S-A0008598097/IXYS-S-A0008597104-1.pdf?hkey=6D3A4C79FDBF58556ACFDE234799DDF0

Best regards,
Jon


Title: Re: Vvr mosfet
Post by: Bieworm on July 04, 2021, 03:07:34 pm
I do not have experience with this particular MOSFET, but I have made a few VVR's for ~20W amps.

According to the full data sheet, whose link I provide below, I looked at the Forward-Bias Safe Operating Area graph (Fig. 12 on page 4).  It showed a DC current limit of about 0.3A at 400V, and 0.2A at 500V.   500V @ 200 mA is 100W.  At lower voltages, of course, the current limit will be higher.  Assuming 25% output efficiency, it appears adequate to me.

The case tab is connected to drain, i.e. the DC input voltage, so the device has to be insulated and heat-sinked properly.

I do not claim to be an expert on the subject of MOSFET VVR's, and I welcome more experienced members correcting any mistake in my analysis.

Full data sheet:
https://4donline.ihs.com/images/VipMasterIC/IC/IXYS/IXYS-S-A0008598097/IXYS-S-A0008597104-1.pdf?hkey=6D3A4C79FDBF58556ACFDE234799DDF0

Best regards,
Jon

Thx, I ordered those: STW13NK100Z
1kV, 13A and 380W.. should be adequate I think.
Title: Re: Vvr mosfet
Post by: vibrolax on July 05, 2021, 09:48:10 am
I looked at the datasheet for STW13NK100Z, and no chart of the Safe Operating Area for DC operation was provided.

Most modern MOSFETs are not intended for operation in the resistive (DC) regime, so it's hard to determine what the current limit is for VVR applications.

The 13A rating only applies for switching applications.  Still, one would hope to operate at DC at 2% of the current limit.

Jon
Title: Re: Vvr mosfet
Post by: Bieworm on July 05, 2021, 11:58:39 am
I looked at the datasheet for STW13NK100Z, and no chart of the Safe Operating Area for DC operation was provided.

Most modern MOSFETs are not intended for operation in the resistive (DC) regime, so it's hard to determine what the current limit is for VVR applications.

The 13A rating only applies for switching applications.  Still, one would hope to operate at DC at 2% of the current limit.

Jon
Ah crap.. I don't know enough on this subject.
Title: Re: Vvr mosfet
Post by: PRR on July 05, 2021, 02:20:06 pm
> I looked at the datasheet for STW13NK100Z

Thanks. (No data-link, no typed-out part number.... what me work?)

That's a high-strung *switching* MOSFET. A polo pony. We want an ox.

The fact that SOA is not plotted for >10mS is a clue. We know the DC SOA is below that, but how much? If we plot the nominal loadline for VVR work we seem to have plenty of clearance. But what I think really kills VVR MOSFETs is that turn-on surge into an empty capacitor. The vertical stub on my loadline.
Title: Re: Vvr mosfet
Post by: Bieworm on July 05, 2021, 11:26:48 pm
> I looked at the datasheet for STW13NK100Z

Thanks. (No data-link, no typed-out part number.... what me work?)

That's a high-strung *switching* MOSFET. A polo pony. We want an ox.

The fact that SOA is not plotted for >10mS is a clue. We know the DC SOA is below that, but how much? If we plot the nominal loadline for VVR work we seem to have plenty of clearance. But what I think really kills VVR MOSFETs is that turn-on surge into an empty capacitor. The vertical stub on my loadline.

Mmm, confusing... bottom line is the thing WILL fail?
Title: Re: Vvr mosfet
Post by: Amplodyte on July 07, 2021, 02:12:30 pm
How about this one:
https://www.mouser.com/datasheet/2/308/1/FQPF9N90C_D-2313782.pdf

Am I reading the SOA correctly to include DC at 400V up to about 500ma for the FQP9N90C version?
Title: Re: Vvr mosfet
Post by: vibrolax on July 08, 2021, 04:03:05 pm
Nope.  At 400V, the dc current limit is under 200 mA.  The voltage and current axes are both logarithmic.

Jon
Title: Re: Vvr mosfet
Post by: Amplodyte on July 09, 2021, 09:40:01 pm
We're discussing figure 9-1 for the FQP9N90C mosfet that I cited, correct? I see the DC line intersect the 400V (4 X 10^2) ordinate close to the 500mA (5 X 10^-1) abscissa. What am I missing?
Title: Re: Vvr mosfet
Post by: PRR on July 09, 2021, 11:51:39 pm
The question seems to be: how bad is the initial current surge? Do you even know?

I've heard of far too many MOSFETs dying in this application to be sanguine. Knowing some about the direction of "progress", I suspect we want OLD DUMB devices, not the latest hot/fast stuff.
Title: Re: Vvr mosfet
Post by: Bieworm on July 10, 2021, 03:31:32 am
I have some old devices where I can look for them oxes 😄
I have purchased some 2W 750V pots, so I can start experimenting.
For the build I'm doing I bought the tube town VoCom G2 controller. Wonder if that will be a good alternative. https://www.tube-town.net/ttstore/kit-tt-vocom-variable-voltage-regulator.html
The Marshall Origin series does attenuation by manipulating G2 , but with a switch instead of a pot. Those amps are doing just fine with that approach. Worth a try ...
I'm building an 18 watt Superlite TMB with the addition of a 1 tube reverb.
Title: Re: Vvr mosfet
Post by: vibrolax on July 10, 2021, 09:24:12 am
We're discussing figure 9-1 for the FQP9N90C mosfet that I cited, correct? I see the DC line intersect the 400V (4 X 10^2) ordinate close to the 500mA (5 X 10^-1) abscissa. What am I missing?

Yes, you read it correctly.  I was looking at figure 9-2; I didn't notice that there were charts for 2 device variants.

Jon