MPP

[Joachim Gerhard]

A bit like this but with less gain.

 

[planet IX]

I was also thinking about to build something similar to this version. I think, the gain formular in the picture is wrong. Should be (R5+R4) /R9.
By making two single voltage deviders to the Opamp +input, offset could be adjusted, right? (a pot in one leg to V- or GND)
R23/C2 could be omitted for two small Cs between the Opamps - input and output, like in Hans' design.

Boris

 

[Joachim Gerhard]

The idea with the two voltages is interesting.

 

[Joachim Gerhard]

I also will build the Kirkwood pre-pre so I can compare.

 

[Hans Polak]

Even more precise, gain is [1+(R4+R5)/R9]

If this pre-pre is meant to be used in front of a MM Riaa amp, why is gain that high?

If going to be used in front of a A/D where a DSP will apply the time constants, be aware that the Vbe differences between the two transistors are temp. dependent but will only cause a few hundreds of mV at the output.
So either caps in case the A/D doesn’t have them already or a servo will do the job.

Hans

 

[Hans Polak]

I also will build the Kirkwood pre-pre so I can compare.

Isn’t the only difference between the two that yours is SE out and Wayne’s version is differential out?

Hans

 

[RCruz]

Joachim,
Since you mentioned a bipolar version, I had already made a bipolar version for myself based on the same design topology but with halve the current in the LTP.
This bipolar version has exactly the same noise and gain specs as the IF3602.

So just for the record, here it is.

Hans
.

I can not make it work as there is always a negative voltage in the collector of Q1 (ccs npn)

Maybe it is due to the model I am using:

*ZETEX ZTX851 Spice Model v1.0 Last Revised 21/1/93
*
.MODEL ZTX851 NPN IS =1.0085E-12 NF =1.0001 BF =240 IKF=5.1 VAF=158
+ ISE=2E-13 NE =1.38 NR =0.9988 BR =110 IKR=5.5 VAR=46
+ ISC=4.6515E-13 NC =1.334 RB =0.025 RE =0.018 RC =0.015
+ CJC=155E-12 MJC=0.4348 VJC=0.6477 CJE=1.05E-9
+ TF =0.79E-9 TR =24E-9
*

Would you share your LTSpice model for the ZTX851 ? I have some in my stock.....

 

[Joachim Gerhard]

His circuit is Hard to read. I think i have to draw it my way to understand it completely.

 

[as_audio]

I can not make it work as there is always a negative voltage in the collector of Q1 (ccs npn)
.....

Post 11569 ? This voltage must be negative, if the input is grounded.
In the case of bipolar input transistors the ccs has to be set up with a
separate reference voltage.

 

[Hans Polak]

I can not make it work as there is always a negative voltage in the collector of Q1 (ccs npn)

Maybe it is due to the model I am using:

*ZETEX ZTX851 Spice Model v1.0 Last Revised 21/1/93
*
.MODEL ZTX851 NPN IS =1.0085E-12 NF =1.0001 BF =240 IKF=5.1 VAF=158
+ ISE=2E-13 NE =1.38 NR =0.9988 BR =110 IKR=5.5 VAR=46
+ ISC=4.6515E-13 NC =1.334 RB =0.025 RE =0.018 RC =0.015
+ CJC=155E-12 MJC=0.4348 VJC=0.6477 CJE=1.05E-9
+ TF =0.79E-9 TR =24E-9
*

Would you share your LTSpice model for the ZTX851 ? I have some in my stock.....

Your model is correct, but Rbb is way too optimistic, that value RB should be modified from 0.025 to 1.5.
This value of 1.5R has been measured and verified.
But it cannot be the reason why your file does not work, it only influences the noise figures.

Here is the version with two 500R resistors instead of 1K in the LTP, and at the same time the resistance between the emitters has been reduced from 7.8R to 3.95R to keep te same overall gain of 61.8dB.
By lowering these resistors values, the A-weighted equivalent noise has been even further reduced to 33.9e-6/61.8dB/sqrt20Khz = 0.20nV/rtHz.
Not bad for transistors costing only 1,-Euro a piece, that don't even need to be matched.


Hans.
.

 

[RCruz]

Thank you again Hans.

I will also try the same circuit with PBSS4350SS instead of the ZTX851... I have a working proto with those transistors working @ 12mA and they are very well matched.

 

[RCruz]

Post 11569 ? This voltage must be negative, if the input is grounded.
In the case of bipolar input transistors the ccs has to be set up with a
separate reference voltage.

Right.... thank you so much

 

[Hans Polak]

Ricardo,
There are no transistors that can even match the noise of the ZTX851/951.
See "The art of Electronics" from Horowitz and Hill.

The one you mention has impressive specs but has no specs on noise.

Hans

 

[RCruz]

Now I understand why my sim did not work... I kept input base to gnd resistors as 220k.

Replacing those with 1k made everything get back to normal.

As I prefer to load my benz cart with 1k5 ohms I need a slightly higher base ref to gnd... will try 2k

 

[RCruz]

Ricardo,
There are no transistors that can even match the noise of the ZTX851/951.
See "The art of Electronics" from Horowitz and Hill.

The one you mention has impressive specs but has no specs on noise.

Hans

I understand hans but those biss bjt come ultra well matched.

 

[Hans Polak]

Now I understand why my sim did not work... I kept input base to gnd resistors as 220k.

Replacing those with 1k made everything get back to normal.

As I prefer to load my benz cart with 1k5 ohms I need a slightly higher base ref to gnd... will try 2k

Two 1k resistors to gnd mean 2k for the cart.
So when replacing the 1k over the input by 6k, you will have 1.5 for the Cart.

Hans

 

[Hans Polak]

There is however an other problem to be solved.
Not too much current should flow into the Cart, a problem that’s not there with Fets.
But there is a simple remedy to solve this with a trimpot.

Hans

 

[RCruz]

Please explain...

There is always some voltage in both inputs... to avoid current flowing through the load those voltages should be precisely equal.... Can we place a pot in parallel with one (or both) resistors from input bases to gnd ?

 

[Hans Polak]

The solution is:
Instead of the two 1k resistors to gnd, use a 2k trimpot with the centre tap via another 1k resisistor to gnd.

The diff input impedance will then be 2K and the CM input impedance ca. 2k.
This of course without the resistor directly over the input.
After having set the 10R pot in the LTP and connection of the servo, this all done with shortened input, look what output voltage the servo shows.

Now remove the input short and turn the 2k input pot until the servo shows the same voltage as with shortened input.

When this is achieved, there is no voltage on the input and no current flows into the Cart.

Hans

 

[Joachim Gerhard]

There is another solution that is used in the Paradise. Here the potmeter is 100kOhm and the resistor is 68kOhm. That way the trimmer does have a smaller effect on the input resistance.