markw4.... you can borrow the Panasonic … it measures thd, thd+n and IM and is easy to use. But you have to make that dreaded trip up and down the canyon 🙂
It measures to -140.
BTW - the Crown dsp is a LOT better than the minidsp from analog in to analog out.
-Richard
It measures to -140.
BTW - the Crown dsp is a LOT better than the minidsp from analog in to analog out.
-Richard
@MarkW4, for the cases you described, differential techniques using arbitrary signals including music, preferably with time-domain averarging, would certainly show much more what's going on than any "standard" test method of the classic analyzers. True sample-synchronous record-while-playback with low drift/jitter is typically required for this to get down really deep into the residual and many other details need to be addressed, uC-controllable special-purpose hardware designed for the particular test, pre- and postprocessing of the signals, things like how to manage clock and gain drifts of the measurement hardware, etc.
As for Richards question...
Refining the design and implemetation of such tests is what I'm working on, for years now and probably for many more to come, a true long term project. Other than that I enjoy working on my (mostly tube) guitar and bass amps.
As for Richards question...
Refining the design and implemetation of such tests is what I'm working on, for years now and probably for many more to come, a true long term project. Other than that I enjoy working on my (mostly tube) guitar and bass amps.
Working out a practical method to measure sound power anechoically on behalf of a leading European audio equipment producer; prototyping dual 18" subwoofer that should weigh in at less than half of conventional boxes; some toying with cheap ES9038 I2S boards to improve them sufficiently for use behind a DSP-based xover. Not a complete list.
What are you using ?
2SK170 BL matched pair for the input ?
LU1014 triode cell for the output ?
Patrick
Scott, if you decide to use the M2's Class-A bias circuit with the 4N35 optocoupler, let me suggest that you spend fifteen dollars and two hours, measuring a few 4N35's from each of several different manufacturers. Lite-On, Fairchild, Vishay, Broadcom/Avago, Toshiba, and Everlight are the main vendors. The 4N35 part number is a jellybean product with extremely loose specs (think 1N4004) chosen to ensure "Parametric Yield Loss equals zero". No opens + no shorts? Ship it.
Probably you want to chose the mfr/fab whose LED Vfwd has the smallest standard deviation, at least that's my guess, since the Pass M2 Class A output stage's bias current is linearly proportional to Vfwd. I predict Vishay will win this particular contest, based on what I remember from my own testing a year ago.
(and of course you want the Class-A bias current in the Left Channel amplifier, to exactly match the Class-A bias current in the Right Channel amplifier, within 0.05 dB, so plan on matching Vfwd when you do the final build.)
_
Last edited:
Several decades back in time... some high speed samplers were borrowed from UC Berkeley Lab....(later returned) … capturing music signals with before and after -- with and without polar coupling cap -- showed differences in the sampled music when visually compared side by side of same music portion visually.
Today, with PC and off the shelf software the difference could be captured also and processed further (like harmonic components etc). Where can a good high speed sampler be had today? Make one, maybe? Or is there a better way.... like the diff measurement W.Jung and J.Curl used for comparing C's and DA. Just further refined and using music signals etc.
THx-RNMarsh
Last edited:
Mark we do see variation in the 4N35 parts. Vishay is never a bad choice.
All matching needs to be to .001 dB anything less causes insomnia for audiophiles.😉
All matching needs to be to .001 dB anything less causes insomnia for audiophiles.😉
Thanks guys but my first try will be quite different and everyone should know by now that I'm not into obsessive matching. 😉
I have stashes of LSK and SK originals. I did quite well with a little Vgs vs Id tracer I built to curve fit JFET's and extract Vp and Idss and keep them ordered on a piece of foam. I just write a Python script to scan the data and pick them for any application. I also don't worry about current sharing of paralleled JFET's. I would drive you guys nuts.🙂
> Laterals has negative tempco. No need for bias control.
++++++++++++++++++++ 111111111111111111111111111 !!!!!!!!!!!!!!!!!!!!!!!!
Only power xsistors I use 🙂
++++++++++++++++++++ 111111111111111111111111111 !!!!!!!!!!!!!!!!!!!!!!!!
Only power xsistors I use 🙂
> I also don't worry about current sharing of paralleled JFET's. I would drive you guys nuts.
So you are paralleing 500x 2SK170's for the output stages ??
😀
Patrick
So you are paralleing 500x 2SK170's for the output stages ??
😀
Patrick
I would be curious to learn how you came to decide on this topology for yourself. Out of the millions of possible ways to make an amplifier for audio/music, you landed there. Was it the simplicity of it? Or ??
I know of one PA design (B.Elliot) that used only 3 transistors and high efficiency speakers which sounded very good/enjoyable. Class A at 3W only.
THx-Richard
Last edited:
Hi
Could I have some advice on how to preserve the leads of my Toshiba JFETs? They are starting to show some pitting.
Should I cover them with solder? What solder and method would you recommend?
I thought of clamping the leads to a heatsink, place it upside down (die down, leads up) to minimize thermal stress to the device while covering the leads with solder.
I appreciate your help.
-Alex
Could I have some advice on how to preserve the leads of my Toshiba JFETs? They are starting to show some pitting.
Should I cover them with solder? What solder and method would you recommend?
I thought of clamping the leads to a heatsink, place it upside down (die down, leads up) to minimize thermal stress to the device while covering the leads with solder.
I appreciate your help.
-Alex
Are you not familiar with Mr. Pass? It's about fun and entertainment. Two channel audio reproduced through speakers in an arbitrary room is always an approximation or maybe better a simulacrum. Accuracy IMO becomes meaningless and what one "likes" or enjoys is all that matters.
Attachments
Sure I had some hearing loss as I get older. Are you not getting older? Last I swept a function gen I was done at 13.5khz. That was in 2013. I'm sure it's a bit worse now. I also have unequal hearing from ear to ear. I can still hear cables, 0.01 caps in parallel with reservoir caps break in, Cartridge loading resistor changes, wire impedance changes on a filament supply line. Need I go on? It's not what you got but what you can do with it.
Oh I forgot; I also have allergies so my ears are popping all for the time from wax deposits. have to clean them with hydrogen peroxide constantly.
With all that I can walk in a room and nail whats going on with a system fast.
Last edited:
How much of that do you figure you can do blind, not necessarily ABX, but just some kind of blind? All of it, some of it?
- Status
- Not open for further replies.