Preamp Freq. Resp. Testing

[guavatone]

I am trying to test my V78. for now i am using a UTC a-11 (10:1) with a level control. For now I am getting results that seem acurate enough but i want the most acurate solution. I am going out of my 96i/o HD to send the test signal. For now, the level must be turned way down, so I think a higher turns xfmr would be better.

I am sure this has been talked about.... getting a clean high Z line level to the input of a 600Ohm Preamp. Any links?

better to use semis for flat response?

I'd like to make a dedicated box for this testing.

I was thinking about using these cheapos:
http://www.mouser.com/catalog/specsheets/XC-600134.pdf
clean enough? or cinemag DI or edcor?

ug, too many choices :?

BTW, I plan to post a Diary of curves for my V78 Mods

:?

 

[skipwave]

http://groupdiy.com/index.php?topic=14468

 

[Samuel Groner]

I am sure this has been talked about...

Indeed: www.groupdiy.com/index.php?topic=14468 [Edit: skipwave was faster...]

Getting a clean high Z line level to the input of a 600 Ohm preamp.

I don't understand your thinking here. You don't want line level to test a micpreamp, you want mic level. You don't want high Z, you want about 150 Ohm, which is relatively low Z and similar to a microphone. And I doubt that the V78 has a 600 Ohm input, although I might be wrong here.

In any case I don't see the need for a step-down transformer if a simple three-resistor PAD as shown in the thread linked above will be the better (lower cost and much higher accuracy) solution.

Samuel

 

[guavatone]

[quote author="skipwave"]http://groupdiy.com/index.php?topic=14468

[/quote]

Uh, can someone make this box for me? It's way too complicated... :shock:
:wink:

yeah, the V78 is terminated(with my current mod) with 400K on the I/P secondary with a 1:20 ratio yielding about 1.1K primary Z

-it's gonna need to be a huge pad so i may go with a bridged-T on this.


thanks for making me feel young again :thumb:

 

[skipwave]

[quote author="guavatone"]-it's gonna need to be a huge pad so i may go with a bridged-T on this[/quote]

The U-pad will be better in this app, since you want a mic impedance looking into the preamp. If you need more attenuation you can raise the 1k series resistors, as noted in Dave's drawing. If you need to reflect a different impedance into the preamp, 50 ohms or something, just re-figure the U-Pad values.

http://www.uneeda-audio.com/pads/

Maybe 7k5 series and 150 shunt is what you need.

 

[CJ]

howabout white noise, or a sweep test-the one that goes from 20 to 20 k anfd back, within a second or djorter?

how about an impulse test-put a spike into the amp, and see what comes out.

if it is a multistage amp. points of interest would be to find out if each stage has the same impulse response, and if not, how do they differ, and what is the effect of combining two similar stages or two dis similar impulse stages, or even four stages, like the v76, then you have a lot more permutations. Some freak over at stanford did a thesis on this, anything for a phd, i guess. interesting reading...

each trsansformer, in and out, can be tested as a pulse transformer, find out what the slew rate is, does a transformer have a slew rate?

good question...

:idea: :roll:

 

[guavatone]

[quote author="skipwave"]
Maybe 7k5 series and 150 shunt is what you need.[/quote]

i remember reading that high series resistors on V72's pads aren't the best. hmm... but it should only reflect Z back to the mic, right?

thanks CJ, I hope to get to the point of interstage and FB loop test. I am a little limited with my test software for now and I am just sine-sweeping with Fuzzmeasure on Mac. Maybe I will look into mac the scope.

Also, so far I changes the ip secondary and that tamed out the 15K boost that was off the charts. Now above 10K is "mr wiggles" but within 6 dB or so.

 

[CJ]

here is a bunch of free software, but the links are dead, maybe you can bak trak and get them:

http://www.audiogrid.com/audio/

 

[guavatone]

different V78 module without mods vs V72

V78 modified: ip primary changed to series with 430K Termination on secondary

I had t o update these unaccurate sweeps

refer to post below.

 

[audiox]

Guavatone, what is the software that you have used for the frequency plots? A freeware maybe? Looks good.

 

[guavatone]

Fuzzmeasure for the mac. Good simple frq resp. in demo. pay for it and I think it does FFT

 

[Samuel Groner]

The low frequency ripple you see on the graphs is a measurement artefact. I quickly played with Fuzzmeasure and it looks like one needs to use a very long sweep time to reduce those to an acceptable level. Not sure how much I'd trust these results.

Samuel

 

[guavatone]

Sam I just figured to get a more accurate graph you need to isolate the signal sweep below the graph. I changed the above graphs to reflect what I did in fuzzmeasure. There must deffinitely be more going on that this simple test isn't seeing.

Anyone got the Neutrik tester around here?

As you can see there's not much aof a change in the V78 with these mods. As far as I can see I have modded everything on the V78 to V72 specs other than the level pot which I keep at low settting giving me 1.3K between suppressor grid/K and gnd. Maybe I am missing something in the o/p xfmr which is labeled V72.

I am wondering what is going on in the V72 after 20K too. It could account for the "airy" quality i like about these.

 

[guavatone]

Time for a drink!

I had a low value cap for the o/p which I didn't realize was 1.5uF and it was cause my freq to rise at 10K and causing a 6dB boost. Replacing the cap with 3.3uF, now it's close to the V72 with a smoother HF curve. I think I'll post a diary soon since I uncovered some key info from doing this.

Anyone else with V72 curves feel free to post please. Age could be changing how they sound.

 

[PRR]

> what is going on in the V72 after 20K too

HIGHLY dependent on source impedance. Which is not cited for either graph. And not known to be equal with/without pad. Or known to be similar to your favorite microphone.

If you like a 3dB rise 16KHz-20KHz, your DAW will do that for you.

Can you make those graphs wider?

 

[guavatone]

[quote author="PRR"]HIGHLY dependent on source impedance. Which is not cited for either graph. And not known to be equal with/without pad. Or known to be similar to your favorite microphone. [/quote]
thanks PRR.

Z out is 50 Ohms feeding a U pad then preamp input z around 1K
the pad is the suggested 7k5 series and 150 shunt with the series resistors on the DA's output side.

[quote author="PRR"]
If you like a 3dB rise 16KHz-20KHz, your DAW will do that for you.[/quote]
We know it's not as simple as doing a sine sweep, but I do know that my DAW will not emulate a V72 preamp with ef804s tubes. It just seems like a starting point to reference what is going on, albeit simplistic.

I am Using some of Ollie's info on V72/V78 units which he says that it's hard to see what's going on without the Neutric tester. I don't have one of those but I would be willing to try a couple of other measurements. Will this be in vain though? Also my lack of knowledge and a really good software app are limiting me a bit here.

[quote author="PRR"] Can you make those graphs wider?[/quote]
I can try, the app is a bit limitted though

 

[CJ]

i think he was being a bit facetious when he said to make them wider.
hey, i learned to spell facetious, did not know it had that t in there.
full props mozilla spell chk.

 

[guavatone]

Sorry PRR. Every post I've read by you has been of a very serious nature so....

Audiovox and Sam:
I found some pretty slick software for mac here.
http://www.faberacoustical.com/products/signalsuite/

 

[guavatone]

[quote author="skipwave"]

[/quote]


I found this input to react strangely with a PAD after it. I had 7.5K series w/ 150R shunt. Then I wanted to see what type of pad my preamp prefered. Long story short using a 600 ohm bridged-T made some strange HF response. I think this PAD may be causing inaccurate measurement of the 2nd PAD shich is the standard ballanced L PAD with 200R Shunt.



Another test I ran was caps feeding a V72 unit. Do you think this HF resp. is a result of the Caps or the caps reacting to the transformer i/p( eg someting more complex)? The caps are Nichicon VX axials 100/63V placed in circuit for Phantom blocking.

 

[CJ]

no, that value cap would only resonate at earthquake frequencies.
light weight caps = high freq
big lytics = loow freq
you get a gut feel just by looking at it.
same with transformers, big = bass, small = tv set IF, Beyer mic in
seat of pants intuition
now look at that lytic, pick it up, throw it against the wall, what does it sound like, boom.
throw a ceramic 0.001 against the wall, it goes tink.
physycal = mental

thats either transformer resonannce from either in or output, and/or shifting of this res point by way of disturbing the circuit with un realistic conditions, ie sig gen instaed of real music, sig gen instead of real impedance expected, output loaded with scope leads that have stray c, use 10:1 when measuring freqs above ten k.