Aleph-X builder's thread.

[Rob Dingen]

Hi Noisfree

Thank's for the schematic.
Take a look at the adjustments from the Sheff UGS (attachment) they work realy wel.
Also the value from R2 and R4 is not the same.

Rob

[noisefree]

Call me Dirk.... :-)

Not really "noisefree" tonight. Have drunken to much redwine! ;-)

There are a lot of possibilies where to put in the trim pots.
The UGS module is ready so I don't wanne change the layout again.
The only thing what I can do in my situation is to raise the values of the jfet loading resistors on the negativ half and reduce the value of the abs. trim pot. But as I meant before, I do not know if the problem is really the pot size itself. When my eyes are good enough I have seen four equal values for the loading resistors on the original.
It is not a problem to use other values - it is more that I need an explanation why the pot is so sensitive.

Here is a photo of the ready module. Resolution is bad - very old digital camera...

Regards
Dirk

[Rob Dingen]

Hi dirk

First for AC The Positive side sees 2K on the collectors the negative side sees 1k5 at the collectors.
Then assume a total of 10ma running.
The adjustment range is from 7.5v to 12.5 volt.
Find out how much you need and change the resistors R4 R5 and the pot P2.
Another solution is to put also 2K in the negative side and P2 50 Ohm and ad a 22 Ohm resistor above P1.
Or make P1 P2 50 Ohm and all four resistors 1k5.
Your module is very nice but it was better to build a proto first

Rob

[Ian Macmillan]

Originally posted by noisefree

The meassured open loop gain is approx. 57dB with 15k input resistors + 15k virtual ground to ground resistors. That’s a bit to much – results in 31dB feedback...
Are there more possibilities than the input-ground resistors to reduce it, and to which value am I allowed to reduce this resistors?

Good to hear you are making progress.

The UGS is essentially a transconductance amplifier and there are not too many ways of changing the open loop voltage gain if the current gain of the ‘mirror’ is required to be a fixed value (a typical value for the latter being about 4). The most obvious control is the output resistor to ground (or power supplies) but reducing this too far will draw too much current from the module. The only other way to my knowledge is to add source resistors to the diff pair (Rs1 – Rs4 in post 1651). This reduces the gain through local feedback or “degeneration”.

I’m not sure how much influence on open loop gain the resistor from virtual earth point to ground has although it will obviously reduce the overall gain by acting as a potential divider with the input series resistor.

Why do you wish to reduce the open loop gain?

Ian.

[Nelson Pass]

Originally posted by Ian Macmillan
I’m not sure how much influence on open loop gain the resistor from virtual earth point to ground has although it will obviously reduce the overall gain by acting as a potential divider with the input series resistor.

It's a good way to "throw away feedback". While the SuperSymmetric
circuit does a good job of improving performance with low amounts
of feedback, it is sensitive to excessive open loop. Resistors to
ground from the summing junctions of the input not only allow
easy control of that, but make the circuit less sensitive to source
impedances.

[Ian Macmillan]

Thanks Nelson. I think I understand the effect but I will have to "play" with some values in a real circuit and observe the change before I understand the precise impact.

Dirk: sounds like you need to reduce your resistors and measure the change. I don't think you need worry too much about going too low.

Ian.

[Ian Macmillan]

I should learn to think more carefully before I reply as there is clearly more complication in the detail than I had assumed. Perhaps Nelson will be good enough to help me out.

In principle I would have thought that “throwing away” feedback would normally increase the closed loop gain of an amplifier given that feedback is generally negative, i.e. less feedback results in more gain, at least until the open loop figure is reached. The effect of a resistor from summing junction to ground must therefore be subtly different.

I’m also missing how such a resistor effects the open loop gain of the amplifier but perhaps this depends on how one defines/measure the latter. I usually define open loop gain as that of an amplifier with its feedback removed, i.e. no resistor from output to input in this case. A resistor from summing junction to ground doesn’t change the amplification factor in this case although it will reduce the overall gain as measured from the input (this being the other side of the series resistor connected to the summing junction) through potential divider action. Is this what was meant by open loop gain in this context?

So what is going on in the case of the AX? The only theory I can come up with goes like this (and is probably completely wrong). The action of the resistor from summing junction to ground works in two ways. First it reduces the feedback factor (by “throwing it away” or attenuating to ground). By itself this would increase closed loop gain except that the resistor also attenuates the input signal through the action of the potential divider it creates with the input series resistor. The net effect is therefore less overall feedback and lower closed loop gain.

The advantage? I’ll hazard a guess that it is the amount of overall feedback rather than open loop gain per se that supersymmetry is sensitive to, hence reducing the feedback in this way also improves its performance.

Am I close or do I have it completely wrong?

Ian.

[Coulomb]

My new job has given me new contacts at R-Theta. A new Group Buy with lower pricing is possible. I will be buying some extrusion to put my Aelphx back on track.

I have started a new thread in Group Buys if your interested.

Now back to your regularly scheduled programming.

Anthony

[Ian Macmillan]

Anyone able to offer an answer to my previous post (1657)?

Ian.

[EUVL]

I'll give it a try.

Assuming we remove the feedback resistors.

The 2 resistors from the summing junction to ground, together with the input resistors, form an attenuator even in open loop. Maybe you're not throwing away much, but you do throw away something. If I am not wrong, you throw away more when the loop is closed.

Treat one half of the AX as an "opamp", put the +input to ground, add the input resistor, the grounding resistor, and the feedback resistor to the inverting input, and work out the close loop gain, etc. The effect of the grounding resistor should perhaps then be more obvious.

(Yes, I did this maths a while back, as I did not quite understand the performance penalties of the AX with single ended input signals.)


Patrick