You need to read the datasheets a bit further beyond the executive summary. Also you could look up the regulator measurements made by Jack Walton in LinearAudio magazine.
So now we are gatekeeping measurements?
You do realize that your whole argument is that if my LNA+ADC are not calibrated anything that I measured is not to be considered?
You do realize that your whole argument is that if my LNA+ADC are not calibrated anything that I measured is not to be considered?
No. I have actually measured both positive and negative denoisers and base my argument on that.
What did you measure? What implementation of those circuits?
What are the arguments on which you defend the VRDN board's performance? Why do you think everything I said is not true?
Do you think it has an optimal implementation of the denoiser? Do you think it offers the denoiser performance gain as it does on a properly implemented LM3x7+ denoiser board?
Else I don't know why you are discussing about it here.
What are the arguments on which you defend the VRDN board's performance? Why do you think everything I said is not true?
Do you think it has an optimal implementation of the denoiser? Do you think it offers the denoiser performance gain as it does on a properly implemented LM3x7+ denoiser board?
Else I don't know why you are discussing about it here.
You are the only one making claims about VRDN board's performance. So far you have not presented any real proof for you claims. I'll leave you to discuss this further with Mark & others.
Are you sure? It has worked perfectly within the spec for everyone that built it with denoiser enabled on both rails? There were no issues with it?
I offered my arguments, and you keep ignoring them. You haven't explained how the creator of the denoiser circuit advised against use of ground planes for the denoiser, advised against putting the denoiser circuit in the high current path, both of which have not been respected in this design.
I offered my arguments, and you keep ignoring them. You haven't explained how the creator of the denoiser circuit advised against use of ground planes for the denoiser, advised against putting the denoiser circuit in the high current path, both of which have not been respected in this design.
I'll take your counsel. Discussion heats up too much for my liking. I'd rather continue when atmosphere is more friendly.
Last edited:
I agree. It may well be that VRDN board's performance could be improved but I do not understand what is the point of accusing Mark for making the board. Elvee's original thread deals a lot with issues many (me included) have faced especially with negative voltage denoisers.
Anyhow the measurements RickRay made show that VRDN offers much better performance than basic LM3x7 regulators. So why the fuss?
When you say increase the level by +3dB, is that the same as +3dBV? I'm trying to go through this process to calibrate the measurement setup I'm using. This LNA should have a little better performance than the LNA in the other thread, I used ZTX951's, but only one way to see.
If you are using REW with V/rtHz you should see the total noise in voltage in the small window at the upper right hand corner. So you should measure a resistor that adds 3dB voltage to this (i.e. multiplied by sqrt(2)). The thermal noise of this resistor is more or less you LNA noise level.
You should also make a basic verification of the scales by measuring e.g. a 1k resistor. That should measure roughly as a straight line at 4nV/rtHz.
If SW displays ie. 100 nV noise level, you should increase that by 1.41 x factor to 141 nV.
https://www.diyaudio.com/forums/equ...-audio-vol-3-spare-boards-30.html#post6432097
Check post #289 as well.
If +3dBv is correct, a 14 ohm resistor created a +3dbV increase. I then plugged it into a resistor noise calculator:
14.0 Ohm resistor at 20.0 °C within 5.0 Hz to 20000.0 Hz frequency band will have :
Noise Spectral Density = 4.760809e-10 V/√Hz or 0.4761 nV/√Hz
Anyone have issues with how I arrived at this figure?
14.0 Ohm resistor at 20.0 °C within 5.0 Hz to 20000.0 Hz frequency band will have :
Noise Spectral Density = 4.760809e-10 V/√Hz or 0.4761 nV/√Hz
Anyone have issues with how I arrived at this figure?
I used ARTA for the measurement above. I will also use REW and bring the results back when I finish that measurement.
No such accusation from me. I thought perhaps the problem was related to the finicky TI 337 chip. Probably I should just give up on the VDRN board and instead use split secondary with stacked 317 based regulator for both pos and neg.
Last edited:
No, it was not you doing the accusations. The ill-behaviour of TI 337s caused oscillations that could be seen by VRDN not reaching regulated voltage. The slightly lower performance in RickRay's measurement was not caused by oscillations. Further measurements are probably needed to find out the real cause.
I don't think the reading I obtained above is correct.
In REW, I shorted the LNA input lead and got a reading of 2.21mV. I was on the V/square root Hz scale, the reading was in the small box on the upper right.
I took 2.21 X1.41 to obtain the +3dB reading of 3.1161mV.
I used an assortment of metal film resistors and the closest I came was 3.17mV with a 68 ohm resistor.
I put that in a resistor noise calculator and obtained the following:
68.0 Ohm resistor at 20.0 °C within 5.0 Hz to 20000.0 Hz frequency band will have :
Noise Spectral Density = 1.049231e-9 V/√Hz or 1.0492 nV/√Hz
I believe this because the LNA is being driven from a 20VAC wall wart. The power supply of the LNA uses half wave rectification, then CRC, then LM317/337 with denoiser.
In REW, I shorted the LNA input lead and got a reading of 2.21mV. I was on the V/square root Hz scale, the reading was in the small box on the upper right.
I took 2.21 X1.41 to obtain the +3dB reading of 3.1161mV.
I used an assortment of metal film resistors and the closest I came was 3.17mV with a 68 ohm resistor.
I put that in a resistor noise calculator and obtained the following:
68.0 Ohm resistor at 20.0 °C within 5.0 Hz to 20000.0 Hz frequency band will have :
Noise Spectral Density = 1.049231e-9 V/√Hz or 1.0492 nV/√Hz
I believe this because the LNA is being driven from a 20VAC wall wart. The power supply of the LNA uses half wave rectification, then CRC, then LM317/337 with denoiser.