20k Rin variant
Update 7 August 2019 - new 20k Rin variant!
Some of you who work mainly with tube equipment have asked for a version with a higher input impedance than the 10k for the stock MK II. Input impedance is a compromise between complexity of the attenuator and the frequency response as was the case with the 100k Rin of the initial version. That 100k required quite a complex calibration for flat frequency response.
The 10k MK II in contrast doesn't need any calibration and is flat to over 1MHz. But listening to you, I build up a unit with 20k input impedance, and tested it. I am happy to report that it has a flat frequency response out to over 600kHz in all settings, without any calibration.
The 20k input impedance version requires some different values for the attenuator resistors. To my surprise, it all fell so nicely together with existing E96 values that 8 less resistors are required, 2 positions on the board have to be shorted and 6 can be left open! Here is the BOM fragment for the 20K Rin variant. All other part values and designations remain the same.
Worst-case attenuation error is 0.1dB in any setting; Rin variation is 19.95k to 20.08k (calculated).
Linear Audio AR MK II Attenuation board Bill Of Materials variant for 20k Rin version only!
Jan
Update 7 August 2019 - new 20k Rin variant!
Some of you who work mainly with tube equipment have asked for a version with a higher input impedance than the 10k for the stock MK II. Input impedance is a compromise between complexity of the attenuator and the frequency response as was the case with the 100k Rin of the initial version. That 100k required quite a complex calibration for flat frequency response.
The 10k MK II in contrast doesn't need any calibration and is flat to over 1MHz. But listening to you, I build up a unit with 20k input impedance, and tested it. I am happy to report that it has a flat frequency response out to over 600kHz in all settings, without any calibration.
The 20k input impedance version requires some different values for the attenuator resistors. To my surprise, it all fell so nicely together with existing E96 values that 8 less resistors are required, 2 positions on the board have to be shorted and 6 can be left open! Here is the BOM fragment for the 20K Rin variant. All other part values and designations remain the same.
Worst-case attenuation error is 0.1dB in any setting; Rin variation is 19.95k to 20.08k (calculated).
Linear Audio AR MK II Attenuation board Bill Of Materials variant for 20k Rin version only!
Jan
Last edited:
Jan - query regarding R48/R49 for the Attenuation PCB BoM. What's the required tolerance? You haven't nominated a recommended part, just a comment and a resistance value.
From looking at the board these resistor should just be straight forward through hole axial resistors, shouldn't they? Why make the comment, 0.4" pitch?
EDIT: Also, what's your take on "low" for temperature coefficient?
Last edited:
Answer to self - board hole spacing. Broke out a ruler and put it up against the board. Came up with 10 mm, which is close enough to 0.4". Got confused by the octopart BoM search function. The have a filter for "terminal pitch" which relates to SMD parts and the rest.
Good job, Jan! It works perfectly fine, I had a lot of fun assembling the AutoRanger MKII.
I really liked the idea with implemented standoffs in the box.
One note during the first power on though - the display could show nothing until you adjust the contrast pot.
It could be recommended using 3mm flat top frost (milky) LEDs on the front panel for OV and UV indication, they look awesome.
I feel like it can be good to shield the SilentSwitcher, and also use some ferrite core clamps for PS cables.
It can be an overkill though. Everything looks almost perfect as is.
I really liked the idea with implemented standoffs in the box.
One note during the first power on though - the display could show nothing until you adjust the contrast pot.
It could be recommended using 3mm flat top frost (milky) LEDs on the front panel for OV and UV indication, they look awesome.
I feel like it can be good to shield the SilentSwitcher, and also use some ferrite core clamps for PS cables.
It can be an overkill though. Everything looks almost perfect as is.
Last edited:
Well done Gordon! Yes frosted LEDs could be nice, really a matter of taste, but why not, they are not critical.
I don't believe that screening the silentswitcher is worthwhile, there is the physical distance, as well as a layout specifically for low radiation.
Did some testing with a spectrum analyzer and I could detect the converter frequency (1.2MHz) with the probe directly at the board. Half an inch away and you see nothing.
But ferrite core clamps on the supply cable might be a good idea, especially on the USB cable to the charger. If I do spectrum analysis I can lower the noise with one or two clamps on the USB cable. But when powered from a Powerbank, I didn't see any difference.
Jan
I don't believe that screening the silentswitcher is worthwhile, there is the physical distance, as well as a layout specifically for low radiation.
Did some testing with a spectrum analyzer and I could detect the converter frequency (1.2MHz) with the probe directly at the board. Half an inch away and you see nothing.
But ferrite core clamps on the supply cable might be a good idea, especially on the USB cable to the charger. If I do spectrum analysis I can lower the noise with one or two clamps on the USB cable. But when powered from a Powerbank, I didn't see any difference.
Jan
Jan, you note over at your web page that:
EDIT: Just to confirm, the power rating of R15 / R25 appears to be no less than 0.5 W. How low can it acceptably reduce to?
Tolerance is spec'd at 0.1%. Any room to relax there?
Note that I fully expect these specs to be "un-relax-able" but I thought that it doesn't hurt to ask.
The BoM notes:
Thanks for to you and Dennis for that intel. I was just starting to scratch my head becauser Mouser told me that the original spec R15/R25 is due for restocking in December.
EDIT: Just to confirm, the power rating of R15 / R25 appears to be no less than 0.5 W. How low can it acceptably reduce to?
Tolerance is spec'd at 0.1%. Any room to relax there?
Note that I fully expect these specs to be "un-relax-able" but I thought that it doesn't hurt to ask.
Last edited:
MKII U3 choices
Hey Jan,
I'm starting to look at stuffing my new MKII board to retrofit my original AR build. I see there are provisions for either NE5534 or AD797 in U3. Could you share with the benefits or reasoning to chose one over the other? Given that the AD797 is 10x as much, do you expect the performance benefit to be worth it?
AD797: https://www.mouser.com/ProductDetail/Analog-Devices/AD797ANZ?qs=/tpEQrCGXCyE0ztKrutblA==
vs
ND5534: https://www.mouser.com/ProductDetai...=sGAEpiMZZMtCHixnSjNA6CumnoLUElGjy/eFGFMqIic=.
Thanks,
Greg
Hey Jan,
I'm starting to look at stuffing my new MKII board to retrofit my original AR build. I see there are provisions for either NE5534 or AD797 in U3. Could you share with the benefits or reasoning to chose one over the other? Given that the AD797 is 10x as much, do you expect the performance benefit to be worth it?
AD797: https://www.mouser.com/ProductDetail/Analog-Devices/AD797ANZ?qs=/tpEQrCGXCyE0ztKrutblA==
vs
ND5534: https://www.mouser.com/ProductDetai...=sGAEpiMZZMtCHixnSjNA6CumnoLUElGjy/eFGFMqIic=.
Thanks,
Greg
Hi Jan.
I have a question regarding my autoranger made by you.
Until now I have only used it for testing a/b amps, but just finished
A stereo class d amp based on iraud9, do you test class d amps with your
Autoranger and what type of low pass filter do you use to dampen the carrier wave?
Mine run 300khz carrier,
I have a question regarding my autoranger made by you.
Until now I have only used it for testing a/b amps, but just finished
A stereo class d amp based on iraud9, do you test class d amps with your
Autoranger and what type of low pass filter do you use to dampen the carrier wave?
Mine run 300khz carrier,
Hi Greg,
It can make a difference if you work with lowish signal levels, with the AR working at +12 or +18dB gain. In that case a '797 will show lower distortion and noise.
Jan
Interesting question. I haven't used it with class D yet. I would at any rate use the balanced input for that.
You can use a filter between the amp output and the bal input, like two resistors in series with each amp output (I guess it is balanced) with a capacitor between them, then on to the AR bal input. I would filter from 30 or 40kHz, which, with 1k resistors requires about 2.2nF. You could put two of those in series for double filtering.
Let us know how it works out.
Jan
I don't see that in my file, but I don't know your real name. Send me a PM with details please.
Jan