Hoffman Amplifiers Tube Amplifier Forum

Other Stuff => Solid State => Topic started by: Fresh_Start on February 27, 2010, 11:50:30 pm

Title: MOSFET buffer
Post by: Fresh_Start on February 27, 2010, 11:50:30 pm
I have one "extra" tube triode in a single-ended amp design and want to add an effects loop.  The insertion point would be between the "master" volume and the PI-like triode stage (i.e. power tube driver).

Kevin O'Connor has a really complicated solid state buffer for an FX loop.  His tube-driven buffer is less complicated, but he says that a MOSFET has too much input capacitance to substitute for the tube triode in the simpler circuit.

Merlin's site shows an AC Coupled Cathode Follower (http://www.freewebs.com/valvewizard/accf.html)  Is it possible to substitute a MOSFET for the tube triode in that circuit?

Can someone point me to a "known good" MOSFET (or other solid state) buffer which does not require a separate power rail?

Ideally, I'd like to use a solid state buffer and O'Connor's overall circuit architecture with the tube triode as the recovery stage.  The series/parallel switch in his design has a lot of appeal.

If it matters, I would not be using any "line level" effects, just pedals.

Thanks in advance for any ideas.

Chip
Title: Re: MOSFET buffer
Post by: FYL on February 28, 2010, 06:21:42 am
Gate/source capacitance isn't really a problem if the mosfet is used as a source follower.

Check http://www.geofex.com/Article_Folders/mosfet_folly/mosfetfolly.htm for some ideas, and don't forget to add gate stoppers (RG hasn't drawn them).

Title: Re: MOSFET buffer
Post by: PRR on February 28, 2010, 05:15:51 pm
> a single-ended amp design

You can steal a mA at 20V from the output cathode, use a JFET.

> Gate/source capacitance isn't really a problem if the mosfet

What's a good small MOSFET for 400V rail?
Title: Re: MOSFET buffer
Post by: Fresh_Start on February 28, 2010, 09:26:49 pm
Mr. Keen suggested this little bugger for a DC cathode follower:
http://www.mouser.com/ProductDetail/Diodes-Inc/ZVN0545A/?qs=sGAEpiMZZMveMCOqFR6qCMipDFy%252b9W8VLXG6ikMZM%2fA%3d

I am truly ignorant when it comes to solid state electronics, despite reading through the section on effects loops in TUT 1 I don't know how many times.  O'Connor's solid state buffer is no doubt great - it just has a high parts count and I simply don't understand all of the circuit.

Could I use a ZVN0545A in place of the tube triode in the AC coupled cathode follower circuit Merlin shows?  (I'll be well below 450 volts on the B+ at that point in the power rail.)

Thanks again,

Chip
Title: Re: MOSFET buffer
Post by: FYL on March 01, 2010, 06:29:07 am
Quote
What's a good small MOSFET for 400V rail?

Zetex has some nice models such as the ZVN0540A (400V) and ZVN0545A (450V), 700 mW diss, 70p CISS. Available from the usual suspects.

http://www.diodes.com/products/catalog/detail.php?item-id=1433
http://www.diodes.com/products/catalog/detail.php?item-id=1434




Title: Re: MOSFET buffer
Post by: FYL on March 01, 2010, 06:49:13 am
Quote
Could I use a ZVN0545A in place of the tube triode in the AC coupled cathode follower circuit Merlin shows?  (I'll be well below 450 volts on the B+ at that point in the power rail.)

A common mosfet is an enhancement device which doesn't conduct with zero volt gate bias, you either need to use it in a DC CF, as RG shows, or add glue.

OTOH, you could use a depletion Mosfet, which conducts by default and is controlled by a negative bias, just like a tube. Supertex has some interesting models, alas sometimes quite difficult to source. Check for instance their DN2540: http://www.supertex.com/pdf/datasheets/DN2540.pdf

Title: Re: MOSFET buffer
Post by: FYL on March 01, 2010, 06:50:28 am
Quote
You can steal a mA at 20V from the output cathode, use a JFET.

Sure, simple and effective.

Title: Re: MOSFET buffer
Post by: Fresh_Start on March 01, 2010, 10:47:50 am
Quote
You can steal a mA at 20V from the output cathode, use a JFET.

Sure, simple and effective.



Can you show me a buffer circuit using a j-fet?

Is this a reasonable part choice?
http://www.mouser.com/ProductDetail/NXP/BF510215/?qs=me8TqzrmIYU2f0hDdp7vzw%3d%3d

I'm going to edumacate myself, starting with http://ist-socrates.berkeley.edu/~phylabs/bsc/PDFFiles/bsc5.pdf

I'll need a regulated low voltage supply since this amp has a fixed/cathode bias switch.  There are plenty of references for that - more than enough in fact :wink:

Thanks for your help guys!

Chip
Title: Re: MOSFET buffer
Post by: FYL on March 01, 2010, 01:20:21 pm


Quote
Can you show me a buffer circuit using a j-fet?

(http://www.muzique.com/images/buff4.gif)

From http://www.muzique.com/lab/buffers.htm

Quote
Is this a reasonable part choice?

Do you like surface mounted components? SOT23 packages are some of the smallest, with 1.4 mm spacing between the legs...
 :huh:

Get some J201's or similar generic jfets in TO-92 package, cheaper and much easier to use.

http://se.mouser.com/ProductDetail/Fairchild-Semiconductor/J201/?qs=sGAEpiMZZMs8pq%2frIcncIDuloSKJeQKQyjKPsc1lhn4%3d

Title: Re: MOSFET buffer
Post by: PRR on March 01, 2010, 03:09:02 pm
> ZVN0545A

As FYL says, "Gate/source capacitance isn't really a problem if the {small} mosfet is used as a source follower." The ZVN0545A's 70pFd "input" is not frightening, and in a Follower that is bootstrapped by roughly Gm/Rl or 500/50K or 1/100, giving 1pFd. Crss adds another 4pFd in Follower configuration. This 5pFd total is about what your wiring stray capacitance will be in any case, much less than the 100pFd of a 12AX7 gain stage or 3 feet of cable.

The MOSFET won't self-bias same as a depletion-mode (vacuum) device, and has lower dissipation ability; do this:

(http://i46.tinypic.com/w0p25d.jpg)
Title: Re: MOSFET buffer
Post by: PRR on March 01, 2010, 04:34:30 pm
> you could use a depletion Mosfet..., alas sometimes quite difficult to source.

Huh. Mouser has LND150 (http://www.mouser.com/ProductDetail/Supertex/LND150N3-G/?qs=sGAEpiMZZMveMCOqFR6qCJgvFtNqkMFqcUttxJa5eGo%3d), 4,000 in stock, under a buck.

Capacitances are truly negigible.

The nominal static zero-bias is ~~1.5mA, just right. However the spread is 1mA to 3mA, a bit broad for thoughtless production. A 1K "self-bias cathode resistor" would bring this near 0.5mA for most samples. Then 100K load would work, and run cool even if B+ were at 500V.

> regulated low voltage supply since this amp has a fixed/cathode bias switch

Fixed-bias does not have the advantages in SE that it may have in P-P.

A simple fix-bias is to replace the cathode resistor with a string of 5V Zeners, adjusted to suitable current (monitored across 1 ohm at the bottom of the string).

The buffer does not need a regulated supply. It may want filtering, but buffer current and voltage is low, 2K and 100uFd 35V ought to be ample.
Title: Re: MOSFET buffer
Post by: Fresh_Start on March 01, 2010, 10:20:00 pm
Many thanks!

Now I have to actually build the amp - without the effects loop or buffer first - and then come back to this.  The important answer here is that it can be done relatively simply.  Those articles FYL linked to from the AMZ Lab Notebook are really helpful.

Assuming that the fixed bias/cathode bias switch turns out to be useless, I can always turn it into a pentode/triode switch.

Chip
Title: Re: MOSFET buffer
Post by: jubal81 on October 01, 2013, 09:22:30 pm
> ZVN0545A

As FYL says, "Gate/source capacitance isn't really a problem if the {small} mosfet is used as a source follower." The ZVN0545A's 70pFd "input" is not frightening, and in a Follower that is bootstrapped by roughly Gm/Rl or 500/50K or 1/100, giving 1pFd. Crss adds another 4pFd in Follower configuration. This 5pFd total is about what your wiring stray capacitance will be in any case, much less than the 100pFd of a 12AX7 gain stage or 3 feet of cable.

The MOSFET won't self-bias same as a depletion-mode (vacuum) device, and has lower dissipation ability; do this:

(http://i46.tinypic.com/w0p25d.jpg)

I'd like to give this a shot. What voltage rating would be safe for the .2uF output cap?