Hi
I know that there is an ongoing thread about the design aspects of the Joshua tree attenuator. I am in the process of building a preamp configued as attenuator -> 12dB buffer. Is the Joshua Tree attenuator good for this application?
I have read that the source sees a varying resistance and the buffer a constant one.... is that right?
What disadvantages are there (if any) to this design over a relay based stepped attenuator configured as a pot?
regards
Rob
I know that there is an ongoing thread about the design aspects of the Joshua tree attenuator. I am in the process of building a preamp configued as attenuator -> 12dB buffer. Is the Joshua Tree attenuator good for this application?
I have read that the source sees a varying resistance and the buffer a constant one.... is that right?
What disadvantages are there (if any) to this design over a relay based stepped attenuator configured as a pot?
regards
Rob
Hi Rob,
The Joshua Tree attenuator has a too low minimum resistance seen by the source : 1.8K. I would rather use this :
Buffer -> attenuator -> buffer.
I'm on building one with an AD815 as output buffer but have not choosed yet the one I will use at input.
Using the same kind of attenuator but with bigger resistor values would allow to get rid of the input buffer.
The one you can find here has at least 9K input R :
http://jos.vaneijndhoven.net
- varying input resistance.
- must be driven with some elaborate programmable chip.
The Joshua tree att. allows for a 0.5 dB step from 0dB to -63.5dB : 128 steps of att. with only 7 relays and 14 resistors.
To perform the same with a relay based pot, you will need 128 relays and 128 resistors !!! And not speaking about the logic control.
I'm not programmer by trade but started to learn PIC programming. It's time consumming and even if my code is somewhat messy, I can now make nice things.
Best regards,
Francis
The Joshua Tree attenuator has a too low minimum resistance seen by the source : 1.8K. I would rather use this :
Buffer -> attenuator -> buffer.
I'm on building one with an AD815 as output buffer but have not choosed yet the one I will use at input.
Using the same kind of attenuator but with bigger resistor values would allow to get rid of the input buffer.
The one you can find here has at least 9K input R :
http://jos.vaneijndhoven.net
Yes.
For me, there are two :
- varying input resistance.
- must be driven with some elaborate programmable chip.
The Joshua tree att. allows for a 0.5 dB step from 0dB to -63.5dB : 128 steps of att. with only 7 relays and 14 resistors.
To perform the same with a relay based pot, you will need 128 relays and 128 resistors !!! And not speaking about the logic control.
I'm not programmer by trade but started to learn PIC programming. It's time consumming and even if my code is somewhat messy, I can now make nice things.
Best regards,
Francis
Attached file are the Jos van Eijndhoven's attenuator results : minimum input resistance 22K, constant output resistance 9K.
Another weak point when using relays, though I don't worry about it, is switching noise. I tried different programming strategies without being able to get rid of it. Even fully ac coupled.
For me, the main reason of using this kind of attenuator, is it's ability to feed a buffer wich has an high input bias current. Output offset will stay stable whatever attenuation used and dc coupling becomes less of a problem.
If fixed and higher input resistance is needed, another possibility is a shunt attenuator : a single serial resistor as high as your source has need for, followed by relays wich will connect different resistors to ground. Thus, the sound will only flow through one resistor. (of the highest quality 😉) But you will still have need for a logic control, unless you limit the step number and use a simple multi step switch to feed the relays.
Best regards
Francis
Another weak point when using relays, though I don't worry about it, is switching noise. I tried different programming strategies without being able to get rid of it. Even fully ac coupled.
For me, the main reason of using this kind of attenuator, is it's ability to feed a buffer wich has an high input bias current. Output offset will stay stable whatever attenuation used and dc coupling becomes less of a problem.
If fixed and higher input resistance is needed, another possibility is a shunt attenuator : a single serial resistor as high as your source has need for, followed by relays wich will connect different resistors to ground. Thus, the sound will only flow through one resistor. (of the highest quality 😉) But you will still have need for a logic control, unless you limit the step number and use a simple multi step switch to feed the relays.
Best regards
Francis
You know when you ask a question and it gets answered in one fell swoop... well that is what you have done. Thank you.
So might you choose the 22K version? Can't a 22K - 1k8 version be made easily?
Best
rob
🙂
So might you choose the 22K version? Can't a 22K - 1k8 version be made easily?
Best
rob
🙂
Hi Rob,
It could surely be made with counters and decoders feeded by push-buttons. Though, getting low switching noise that way won't be an easy task. For me, it appears to be more simple with a PIC.
Sorry not to be more helpfull, I'm just an amateur.
Best regards,
Francis
It could surely be made with counters and decoders feeded by push-buttons. Though, getting low switching noise that way won't be an easy task. For me, it appears to be more simple with a PIC.
Sorry not to be more helpfull, I'm just an amateur.
Best regards,
Francis
Hi
I'm thinking of building a Joshua tree with the 9K output impedance and 22K input impedance. What if any problems are there with a high output impedance? The Tree will have an AD815 preamp after it.
thanks
Rob
I'm thinking of building a Joshua tree with the 9K output impedance and 22K input impedance. What if any problems are there with a high output impedance? The Tree will have an AD815 preamp after it.
thanks
Rob
- Status
- Not open for further replies.