Front-end transformer isolation and filter circuit analysis

[helgrind]

Hello,

I'm working to replace a pulse measurement system (from a capacitive pickup) which includes the following isolation and filter stage, but would like to understand a little about the background first. Unfortunately the designer is no longer at the company and there are no design records available. I've been having trouble simulating it in spice, and I wonder how important the transformer impedance is in the response. Note that it is a 9:1 turn ratio.

I have simmed the input LPF successfully, but afaik the bandwidth of the second stage should be around 20 MHz but I am getting a much narrower notch more like 20 kHz, just from the section before the op-amp. I'm also confused that in the second stage, the LPF appears to be before the amp and the HPF after. I thought for an active BPF you normally put them the other way around for compensating nonlinear effects? Unless it's because of impedance reasons here?

Finally, I wonder why the isolating transformer has the caps hung off the grounded side rather than in series with the main output, before the inductors. Would there be an electrical reason for this or just perhaps due to layout reasons?

Any help in analysing this would be much appreciated. The schematic excerpt is at the link below as a png.

Thanks,
Laurence

https://drive.google.com/file/d/1VrwqKQ9H0Z4hRwcMSuqyLJRXNm-pTHI3/view?usp=sharing

 

[D.A.(Tony)Stewart]

This design appears to focus as an IF amplifier. Although the front end is a LPF with some peaking then Box4 has a transformer coupled LPF with high Q peaking but not terminated by 51 Ohms ,rather in series with Vin+ which is fairly high impedance. ,then a low Q BPF with a trimmer. This is a very high BW Op Amp which is not an active filter, just an amplifier with high gain.

 

[helgrind]

Thanks Tony for your reply. I have a few further questions if you don't mind please:

1. Could I ask why you inserted the additional 10pF shunt cap on the left? Is it to cover board capacitance?
2. Where did the 35pF and 2.7k load termination come from? I saw 1k between Vhi and Vlo for the log amp input spec.
3. What's the reason for the capacitive ground coupling on the transformer output and would it matter if that was on the other side of the winding?
4. What's the reason for the 17 MHz LPF on the Vlo input only on the log amp?

Thanks in advance!

 

[FvM]

Transformer parameters surely affect transfer function. Could try S2SPICE tool to generate a realistic model.Turns ratio is 1:3 by the way, impedance ration 1:9.

Several circuit details look arbitrary and not well considered, e.g. the mentioned capacitors at transformer secondary. Effect is reduction of common mode suppression, depending on transformer parasitics and actual common mode noise. Also IF amplifier driving impedance is too high, adding out-of-band noise floor to log amplifier output.

All-in-all, I doubt that it's a good redesign template. Circuit behaviour should be nevertheless analyzed as a reference

 

[D.A.(Tony)Stewart]

the series caps can be moved. I could not show the transformer.