(Just made this thread so I can keep my progress and all my questions together in one place here with respect to the preamp I will be building over the next month or so.)
I just ordered the B1 buffer board with set of 4 matched transistors from Pass Labs DIY store -- excited!
I just now realized that the circuit calls for linear taper pots (shame I just bought a 100k and a 10k alps RK27 pot -- which are log). (Reason I bought the log pots is I thought I was going to go with a passive pre, but I need to split the outputs so one can go to my powered sub and decided --after getting some help from here -- that a buffer would be the best way to do this.)
So what make / model linear taper pot should I get and from what supplier?
I just ordered the B1 buffer board with set of 4 matched transistors from Pass Labs DIY store -- excited!
I just now realized that the circuit calls for linear taper pots (shame I just bought a 100k and a 10k alps RK27 pot -- which are log). (Reason I bought the log pots is I thought I was going to go with a passive pre, but I need to split the outputs so one can go to my powered sub and decided --after getting some help from here -- that a buffer would be the best way to do this.)
So what make / model linear taper pot should I get and from what supplier?
Last edited:
In Nelson Pass's 2008 article on the B1 buffer, he mentions he uses the Axon 10 UF metallized polypropylene caps. I checked a few places and they are all out of stock .. many of the other sizes are available but 10 uf in particular is not even mentioned.
Any good substitutions? And where I should buy them?
Any good substitutions? And where I should buy them?
Thanks I'll get my Axon's from there.
Another question: Would one really notice that much of a difference in the B1 using Vishay resistors (rn55d) vs bulk 1/4 watt metal film resistors from china ($5 I paid recently for qty 600.) Trying to keep the costs down.
Another question: Would one really notice that much of a difference in the B1 using Vishay resistors (rn55d) vs bulk 1/4 watt metal film resistors from china ($5 I paid recently for qty 600.) Trying to keep the costs down.
cmf55/rn55 are not really expensive.
It's hard to say whether you would hear a difference. It would depend on the quality of the parts. At least measure them to ensure the values are correct.
It's hard to say whether you would hear a difference. It would depend on the quality of the parts. At least measure them to ensure the values are correct.
Read Pass and/or Borbely and learn how to measure Idss.
Check your jFETs for Idss.
For test equipment you can get away with just:
a 9V battery (nearly fresh so that the voltage does not droop while you are testing your jFETs).
a 10r metal film 1/4W resistor
a DMM set to 199.9mVdc
A 12mA Idss will measure ~120mVdc and the Vds will be ~9V-0.12V = 8.88V
A 6mA Idss will measure ~60mVdc and the Vds wil be ~9V-0.06 = 8.94V
When you are done and maybe repeated the measurements two or even three times, you select the devices so that you have two pairs that are very close for Idss.
Use one pair in each channel.
Don't solder them in.
Solder in SIL sockets.
Plug in the jFETs and measure the Vgs of the "upper" device.
Vgs should be very close to zero mVdc
Once you have proved you have good pairs, then you can elect to solder the lead outs to the SIL sockets.
Check your jFETs for Idss.
For test equipment you can get away with just:
a 9V battery (nearly fresh so that the voltage does not droop while you are testing your jFETs).
a 10r metal film 1/4W resistor
a DMM set to 199.9mVdc
A 12mA Idss will measure ~120mVdc and the Vds will be ~9V-0.12V = 8.88V
A 6mA Idss will measure ~60mVdc and the Vds wil be ~9V-0.06 = 8.94V
When you are done and maybe repeated the measurements two or even three times, you select the devices so that you have two pairs that are very close for Idss.
Use one pair in each channel.
Don't solder them in.
Solder in SIL sockets.
Plug in the jFETs and measure the Vgs of the "upper" device.
Vgs should be very close to zero mVdc
Once you have proved you have good pairs, then you can elect to solder the lead outs to the SIL sockets.
Last edited:
Just ordered 1 lb of Kester 63/37 .031 solder. Hope I didn't make a mistake by not ordering Cardas Quad Silver instead. Can one really hear an audible difference? I read that particular Kester solder makes a very fine solder joint.
I usually use quad core Ersa fluxed eutectic solder.
Gives very nice solder joints.
Don't worry about Silver solder
Last edited:
AndrewT, thanks. I'll test them like you say. Question regarding the SIL and the JFETS. Should I just not solder the transistors at all? I.e. just plug into the SIL and leave alone? I could de-oxit them like people do with tube pins and tube sockets from time to time. My thinking is the heat from the soldering iron could mess with the transistor? (like permanently alter it even with relatively small amounts of heat) ?
I have some that are just inserted into SIL sockets and after a very few years of use have not heard any effects, but I have not gone back and checked to hear/measure if socketed is different from soldered.
Does it get better or worse after 10years? 20years?
You will hear many recommend always solder.
You can buy a little aluminium thermal clamp to help keep the junctions cooler while prolonged soldering/ desoldering is in progress. I have but rarely use.
There are many times I regret having soldered in To92 devices especially when they were specially selected for that location.
The B1 can have a +ve Vgs, just like a DCB1 can have a +ve output offset. This tells you that the jFET is biased above Idss. very slight high bias does not damage the device. But high level signal will take the Id well above Idss and that is where gate leakage and other effects can dteriorate performance.
Ideally you don't want +ve Vgs for an Nchannel jFET. We normally run them @ 30% to 90% of Idss, B1 & DCB1 are exceptions and run the Id @ ~100%Idss.
Does it get better or worse after 10years? 20years?
You will hear many recommend always solder.
You can buy a little aluminium thermal clamp to help keep the junctions cooler while prolonged soldering/ desoldering is in progress. I have but rarely use.
There are many times I regret having soldered in To92 devices especially when they were specially selected for that location.
The B1 can have a +ve Vgs, just like a DCB1 can have a +ve output offset. This tells you that the jFET is biased above Idss. very slight high bias does not damage the device. But high level signal will take the Id well above Idss and that is where gate leakage and other effects can dteriorate performance.
Ideally you don't want +ve Vgs for an Nchannel jFET. We normally run them @ 30% to 90% of Idss, B1 & DCB1 are exceptions and run the Id @ ~100%Idss.
Last edited:
I used Obligato 10uFs in my build. The type of solder makes very little difference, certainly not audible. But I am in the camp that the components should be soldered rather than in sockets. The semiconductors can easily tolerate the heat involved in soldering for a few seconds without any detrimental affect (the actual time is specified in the datasheets and is usually about 2 seconds), just make sure that everything is shiny clean before you solder.
Attachments
Last edited:
Thank you. There are so many options for "audio grade" metallized polypropylene 10uF caps. I guess it doesn't really matter which one I pick? THey are all about $5-10 each.
The DC blocking capacitor does not generate distortion in the audio passband because the capacitor does not act as a filter in that audio passband.
The DC blocking capacitor must pass the audio signal and preferably not alter the phase of the audio signal. This requires that the Filtering effect of the DC blocking capacitor be at least three octaves below the audio passband.
I and many other suggest the filter created by the DC blocking capacitor be around a decade below the wanted passband. Some will suggest even farther away to reduce the phase effect just a little bit more.
A 10uF DC blocking capacitor feeding a 10k input impedance creates a 100ms RC filter that has an F-3dB ~1.6Hz, i.e. more than a decade away from the 20Hz lower audio passband limit.
That 10uF will have virtually no audio signal across it and thus virtually no distortion of the audio signal passing through it.
You can use an MKT, or any of these plastic film capacitors, in 50V to 100V rating.
If you temporarily use an input impedance of 100k then your DC blocking capacitor can be reduced to 1uF.
Now you can afford to experiment with many different dielectrics and find out if you can hear any difference. Then you can measure to find out what that difference is and from that find out why some sound different. But I don't think you will hear any difference! Except some, especially loose wound (very expensive hand wound types), may suffer from microphonics. That will be audible, but certainly not better.
The DC blocking capacitor must pass the audio signal and preferably not alter the phase of the audio signal. This requires that the Filtering effect of the DC blocking capacitor be at least three octaves below the audio passband.
I and many other suggest the filter created by the DC blocking capacitor be around a decade below the wanted passband. Some will suggest even farther away to reduce the phase effect just a little bit more.
A 10uF DC blocking capacitor feeding a 10k input impedance creates a 100ms RC filter that has an F-3dB ~1.6Hz, i.e. more than a decade away from the 20Hz lower audio passband limit.
That 10uF will have virtually no audio signal across it and thus virtually no distortion of the audio signal passing through it.
You can use an MKT, or any of these plastic film capacitors, in 50V to 100V rating.
If you temporarily use an input impedance of 100k then your DC blocking capacitor can be reduced to 1uF.
Now you can afford to experiment with many different dielectrics and find out if you can hear any difference. Then you can measure to find out what that difference is and from that find out why some sound different. But I don't think you will hear any difference! Except some, especially loose wound (very expensive hand wound types), may suffer from microphonics. That will be audible, but certainly not better.
Last edited:
I plan on using the 20k alps rk27 pot for the B1 I guess. So still go with the 10uf or should I adjust that down?
- Status
- Not open for further replies.