100Hz square preamp responce help

[GoranB]

Any suggestions how to improve the responce of this preamp at 100Hz square signal?

[Alastair E]

Why?--Whats wrong with the square as it is....?

Whats with 'Only Allen Bradley' resistors too...?
--Bit 'pretencious' IMO

Personally, IMO, Its a pretty basic scheme, a gain-stage and cathode-follower, and will be quite load imp dependent.....

There's much better schemes out there....

What load you running it into?

[GoranB]

I have PF99 SE A class 3A per chanell.
After removing the decoupling catode capacitor and ressistor and putting the red led's for gain stage catode biasing i achived some improvement.
It look like this.

[kevinkr]

Please post a capture of a 100Hz square wave loop back test through just your data acquisition hardware.

What is the load impedance you are driving with the output of this thing? What does it look like into higher impedance loads? Is the response affected by the value of that coupling capacitor? If you are driving an 8 ohm load it is unlikely that the coupling capacitor has sufficient capacitance to pass the 100Hz square wave undistorted because the time constant formed with the load and source impedance is too short, this means you need more capacitance or a higher load impedance. Or you can ignore it if the bass performance is sufficient with your headphones since music isn't square waves.

See this, it may help you to visualize what is going on better even though it relates to transient vs steady state behavior in a series RC network:
Technical Paper #1: Square Wave Response of a Series RC Circuit

FWIW I have never understood the mystique around Allen Bradley carbon comp resistors, they are lousy resistors compared to almost anything made these days. Lots of excess noise (no I'm NOT talking about Johnson noise) when you run dc current through them and a significant voltage coefficient which is fine if you like unnecessary distortion generators in the signal path.. (These resistors can be up to 20dB noisier than a good NI WW..) Two watt versions are admittedly a lot better than the 1/2W in both respects, but still.. A nice non-inductive Mills WW has neither of these faults, nor do most metal films to an appreciable extent.. A matter of taste I suppose..

[Yvesm]

And do not forget to set your 'scope in "DC"

[GoranB]

A capture of a 100Hz square wave loop back test.

[GoranB]

[QUOTE=

FWIW I have never understood the mystique around Allen Bradley carbon comp resistors, they are lousy resistors compared to almost anything made these days. Lots of excess noise (no I'm NOT talking about Johnson noise) when you run dc current through them and a significant voltage coefficient which is fine if you like unnecessary distortion generators in the signal path.. (These resistors can be up to 20dB noisier than a good NI WW..) Two watt versions are admittedly a lot better than the 1/2W in both respects, but still.. A nice non-inductive Mills WW has neither of these faults, nor do most metal films to an appreciable extent.. A matter of taste I suppose..[/QUOTE]

Im not using Allen Bradley, i have ordinary cheap ressistors. Output capacitors are Siemens. Honestly i have never heard how Allen Bradley sounds. But I dont believe that there is some magic. Yes it could be a matter of taste

[GoranB]

Originally Posted by Yvesm

And do not forget to set your 'scope in "DC"

Im not using a scope but software. In loop mode the sguare is ok.

[GoranB]

Originally Posted by kevinkr

What is the load impedance you are driving with the output of this thing? What does it look like into higher impedance loads? Is the response affected by the value of that coupling capacitor? If you are driving an 8 ohm load it is unlikely that the coupling capacitor has sufficient capacitance to pass the 100Hz square wave undistorted because the time constant formed with the load and source impedance is too short, this means you need more capacitance or a higher load impedance. Or you can ignore it if the bass performance is sufficient with your headphones since music isn't square waves.

Upppsss.. i forgot the load ressisor on the output of preamp, the input ressistance of PF99 should be 47K. So, to test the preamp it means i should put a 47K on the output?

[kevinkr]

Originally Posted by GoranB

Upppsss.. i forgot the load ressisor on the output of preamp, the input ressistance of PF99 should be 47K. So, to test the preamp it means i should put a 47K on the output?

You need to place a load on the output representative of what you are driving with it, if not headphones then whatever the input impedance of the device you normally drive should be used.

You might want to take a look at your power supply as well - if you are driving anything above 1K input impedance the square wave response at 100Hz should be pretty close to perfect, if it's not you need to look at the coupling cap first - do you have the 200uF shown or something less? Then look at the supply and make sure you don't see a 100Hz square wave on the supply - if you do more supply capacitance is needed. You should use a scope for this or make sure your data acquisition hardware is designed for high voltage use.

One other thought is that if this is being used only as a line stage you can completely remove the second stage as the rp of the 5842 is only a little over 1K so this should be fine for driving a power amplifier directly. I'd use a 1.0 - 2.2uF coupling cap which should be sufficient for an input impedance of 47K. Should sound better..