IRLB3813 Triode-curve

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The design of the amp shown on the previous page as I understand it is greatly influenced by Mr Pass work, I guess that's a given for most of you here, but the choise of going with IRLB3813 in this case is Cirlomanens idea, I have not seen it applied in amps before at least, but even so I know the choise was based alot on seeking the triode characteristic that seems so very appealing.

Going for simplicity in the extreme (my fault, I naged him into it) without resorting to bulky and expensive transformers and depletion mode fet's the design he finaly presented me with (but with a schematic dumbed down by me) can be seen here :

Martinsson's Blog - JRAMP

Bare in mind the citcuit schematic was made by a total newcommer to this side of the speaker terminals, me, and as such I did not use standradized component symbols, but I think you get the hang of it anyway, note that the source resistance is changed from 0,5 to 0,75 Ohm.

So the reson for it not being enhanced by cascoding or using a capacitance multiplier in the PS etc. can be contributed to my limitations as a first time out'er in this field, I wanted it to be as simple as possible to build, and Circlomanen did a great job in designing just that, a simple fully functional amplifier platform for me to improve upon and add to as I learn along the way.
 
Last edited:
Since International Rectifier has released a new series of low voltage switch fets, I spent some quality time studying their datasheets.

The IRFP740 part seems very interesting.

http://www.irf.com/product-info/datasheets/data/irfp7430pbf.pdf

IRFP7430pbf-kurvor.png

You have to interpolate the curves down to to lower voltage Vgs and a device temperature of 60 degrees Celsius (Tj=60C). Just follow the general trend.

I think it is a very interesting device. A cheap way to get triode-like curves without trying to get rare 40 year old Japanese devices.

http://www.irf.com/product-info/datasheets/data/irfp7537pbf.pdf

There is a 60 volt device with similar curves. Also nice looking curves. compared to almost any other normal switch-fet.

Cheers,
Johannes
 
Last edited:
jmf-kurvor.png

Compare the curves from a SIT published by Nelson Pass in his "Introduction to Static Induction Transistors" FIRST WATT SIT INTRO
with the curves from the IRFP7430 part. The IR part might not be as linear and nice sounding as a Nelson Pass SIT1, but I prefer playing with something with much more interesting curves then the normal IRFP-XXX (pentode-curves) but available now for a nice price, then dreaming about stuff that is almost impossible to buy.

Look at the general trend of the curves. Sadly the IRFP7430 is not a depletion-mode device, but this does not effect the sound character.

I guess International Rectifier has been inspired by SIT technology in developing these very fast, rugged and powerful devices. It is a switch fet, and they should be as fast as possible. Every 1/10ths of a percent increased switching performance counts in the very competitive switch fet market today (solar energy, electric cars, battery powered devices and tools, etc etc). The SIT is known to be the fastest solid state device for switching.

Cheers,
Johannes
 
The IRLB3813 was the first low voltage, high current and high transconductance device I tried.

The IRLP3034 is a much better transistor for single ended simple class A amps. TO-247, more "triodity" in the curves, 40 volts Vds voltage-rating. It sounds much better then the IRLB3813 and seem to have the same character over a much wider range of voltages and currents.

The IRFP7430 is a wild and strange transistor. In a simple no feedback Zen like amplifier with 13 volts Vds and 1.7 ampere current it is like a charging rhino on steroids and crack. It is brutal.
You have to hear it to understand. It is not hifi. It is just to much of everything including distortion.

In the right application (phase-splitter amp) it can sound wonderful. I guess it needs a lot of feedback to tame it and domesticate it enough to something resembling a normal mosfet.

These are fun devices. They do not sound similar to the normal IRFP044-054-140-150-240-250 verity Nelson Pass recommends for the Zen projects.
They sound much more "large", alive and vivid, with a whole other presence and drive. Because of their large transconductance I guess they behave like many paralleled IRFP-150 in the very powerful dynamic and controlled bass-reproduction.

Cheers,
Johannes
 
Is there any efficiency to gain?

The efficiency is set by the inductor, not the Mosfet, but there is probably a healthy reduction of distortion and/or a different character to the amp.

With an IRFP7430 you can probably get a more vivid and "forward" character that emphasizes drive, attack, energy and dynamics in the music.

All this is very fun to experiment with, since a Triody device like the IRFP7430, IRFPS3810 or the IRLP3034 is better suited to "working the load line" as Nelson Pass writes about in his SIT and Triode articles.
 
I would never have seen a triode in those curves!

I am also not seen any triode in that curves. but I am sure that circlomanen amp will get vfet like sound. Not because of this IRLB3813, but due to feedback / bias resistors R5 and R4.
Any pentode device will get triode like curve using a feedback resistor from drain to gate. Very easy to implement in mosfets due to high input impedence. Slope of the output curve is defined by ratio of feedback resistors.Also take care of input capacitace of the device when choosing r5 and r4.

Similarly you can convert a triode device to have pentode like curves.
 
Why is these curves not Triode-like?

Even Nelson Pass calls them "Triody".
They sound very similar to a Triode.

I know you can get a Triode like sound with Schade feedback, but with something like the IRFP7430 or IRFPS3810 you don´t need or want Schade feedback or any other kind of feedback except for some resistance between source and ground.
 

Attachments

  • NotTriode.jpg
    NotTriode.jpg
    230.3 KB · Views: 395
Hi Circlomanen,

I am accepting that small slope in the curve but not significant like a triode and your amp is not operating in that circled area. The circled area have current more than 20 Amp, and your amp is biased @less than 1.5A , not clear from curves.

My point is that your amp is more schade than triody, because of R4 and R5. It will help to bias the transistor as well give feed back,both AC and DC. DC feedback will help the amp to get more stable

If you filter the AC component from the feed back you can remove schade effect. Split R5 220 k resistor in to two. one 100k and another 120k connected in series and decouple ac component using a 10uf connceted between junction of 120k/100k resistors and ground.Can you try it and let us know the result.

Regards
Joshvi
 
Triody3810.jpg

Better curves.

The IRFP7430 and IRLB3813 does have a Triody area at low currents too, even though it is not shown in the spec sheet.

kindofTriode.JPG

Here is from Nelson Pass article Zen Variations 8.
http://www.firstwatt.com/pdf/art_zv8.pdf

The IRFPS3810 is much more Triody then the Lovoltech which was Triody enough for Nelson Pass to create two articles and a project with.

A WE 300B might have more "Triody" curves, but a IRFP7430 costs a few dollars and can drive a loudspeaker directly without a step down transformer.

I would like to have a box full of 2SK77, but they are kind of rare, so I have to do with what is easily available, and the IRFPS3810 sounds much better then a Schade-feedbacked IRFP150. I have tried both with bastode-circuits and with transformer based Schade feedback, and I prefer the sound character from IRFPS3810 or IRFP7430 over any Schade feedback arrangement.
 
Member
Joined 2000
Paid Member
Hi Circlomanen,

Have you been working with any new "triody" type mosfets lately?
From my notes, I have the following. Which would you recommend trying first?

IRLB3818 (good)
IRLP3034 (better)
IRFP7430 (wild)
IRFPS3810 (?)

Thanks,

Vince
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.