AMZ MOSFET Boost - some observations

[davebungo]

I've been getting really (too) interested in the AMZ MOSFET boost circuit recently and have been doing a bit of theory and PSpice modelling.  Alas, I haven't tried out any suggested mods yet so I should apologise for this firstly in case someone tries them and it backfires a little.

Anyway, as I cannot post an image, you should refer to the article on Jack Orman's AMZ site.  The main thing that strikes me is that the circuit in its quiescent state is biased in such a way that VDS is about 2.2V.  It's easy to work this out because if you have about 5.6V at the drain (roughly as Jack suggests biasing it) then this means ID is about 1.25mA and this in turn puts the source at about 3.4V.  (5.6 - 3.4 = 2.2)

This means that it doesn't have to swing too far downwards before it bottoms out.  It would be helpful here to be able to show the output characteristic of the BS170 and the PSpice sim I have done so you can see my point.  

So, what of all this?

Well if you try the following values:
R2=240K (was 62K) (tweak to get about 4.5V at the drain)
R4=4.3K (was 2.7K)
R5=470 (was 2.7K),

and leave everything else as it is then VDS is now at about 4V and has more room to swing before the bottom of the output waveform is clipped off due to the FET going into pinch off.  The maximum gain of the circuit is increased to about 34dB (was 31dB I think) but more importantly, the output is able to swing more closely to the full supply than it could before.

This should lead to more headroom and more clean boost before the on-set of distortion which will be more symmetrical.  Note, I am not trying to say this will sound any better - it's just an observation and perhaps worth trying out.

On the down side, it does mean that you can't really use it as a unity gain buffer (without switching in additional resistance in series with R5).  (When R5||R6 (or Rs) is significant (>100 ohms), the gain of the circuit is roughly R4/(R5||R6) or (Rd/Rs))

If anyone is interested in any way at all I would be happy to post more detail e.g. the design equations involved and for anyone who is really, really interested I can recommend "Microelectronics by Millman Jacob, Grabel Arvin".  It does a great treatment of both small signal BJT and FET circuits.

[MartyMart]

That sounds interesting Dave, for R5 you mean 470 ohms, correct ?
You could just use an spdt switch to add in resistance for unity gain, or use a dual ganged pot to control things there and the increased headroom could be great  :wink:
Have you not built one to try that out ?

Regards,
Marty.

[davebungo]

Yep, 470 ohms.

No I haven't modded one yet (yes I know I should do really before I start recommending changes, but I was so interested in this circuit that I couldn't wait to post what I thought - perhaps prematurely).  I think the changes should be relatively safe and they're simple enough to try out and reverse if you don't like the sound.

BTW You should get about 18dB of gain even with R6 at maximum resistance so it will boost by default rising to 34dB of gain at maximum boost.  Remember to tune R2 though to get roughly 4.5 to 5V at the drain of Q1 before you apply any signal.  It should sound cleaner at moderate boost if I'm correct but as I have stressed I haven't proven this in practice and stand to be shot straight out of the water (and deservedly so, maybe).

[seanm]

I've been getting really (too) interested in the AMZ MOSFET boost circuit recently and have been doing a bit of theory and PSpice modelling.  Alas, I haven't tried out any suggested mods yet so I should apologise for this firstly in case someone tries them and it backfires a little.

I couldn't find the suggested mods on the site (except for the unity gain). Maybe the lost hour has dulled my eyes.

Would the same changes apply to a 2N7000? Have you tried modelling a 2N7000 rather than the BS170?

[davebungo]

I couldn't find the suggested mods on the site (except for the unity gain). Maybe the lost hour has dulled my eyes.

Would the same changes apply to a 2N7000? Have you tried modelling a 2N7000 rather than the BS170?

I'm referring to the mods suggested at the start of this thread - sorry if this was confusing.

I haven't modelled the 2N7000 but the data on both devices is very similar so I wouldn't expect a problem.  The Vishay pdf has both devices on the same datasheet.
http://www.datasheetcatalog.com/datasheets_pdf/B/S/1/7/BS170.shtml

[seanm]

I'm referring to the mods suggested at the start of this thread - sorry if this was confusing.

I haven't modelled the 2N7000 but the data on both devices is very similar so I wouldn't expect a problem.  The Vishay pdf has both devices on the same datasheet.
http://www.datasheetcatalog.com/datasheets_pdf/B/S/1/7/BS170.shtml

Great. The 2N7000 that I am using (Fairchild) shows a diode on the schematic symbol. I assume this means I can leave out the Zener?

[davebungo]

I'm referring to the mods suggested at the start of this thread - sorry if this was confusing.

I haven't modelled the 2N7000 but the data on both devices is very similar so I wouldn't expect a problem.  The Vishay pdf has both devices on the same datasheet.
http://www.datasheetcatalog.com/datasheets_pdf/B/S/1/7/BS170.shtml

Great. The 2N7000 that I am using (Fairchild) shows a diode on the schematic symbol. I assume this means I can leave out the Zener?

No, the diode in the package is across the drain and source and doesn't protect the gate-source junction and is (I think) meant to offer some protection in inductive switching applications.

[bwanasonic]

This should lead to more headroom and more clean boost before the on-set of distortion which will be more symmetrical.

Have you ever exceeded the headroom of the AMZ mosfet in actual practical use? It's never come up in actual use for me. I think there was thread on running it at higher voltage for increased headroom.

Kerry M

[davebungo]

This should lead to more headroom and more clean boost before the on-set of distortion which will be more symmetrical.

Have you ever exceeded the headroom of the AMZ mosfet in actual practical use? It's never come up in actual use for me. I think there was thread on running it at higher voltage for increased headroom.

Kerry M

I understand what you're saying, but I'm simply suggesting a possible way to achieve more headroom should you need it.  There have been times when I have wound my MXR micro-amp to full tilt in moments of madness but it sounded pretty good through a JCM800  :wink:

[bwanasonic]

I'm simply suggesting a possible way to achieve more headroom should you need it.

Cool, I was just curious.  

Kerry M