Transistors, not the 3055, but all the BC184's. I used Toshiba 2SC3329BL's.
Renesas 2SC2545 is also good.
No one likes noise, we can agree on that.
Thanks to the Mr.Williams info, posted months ago about this, I did and the improvement was striking in the noise-floor. Cheap (*1) and fast, then I stop "esoterics" moddings, for me pre73 it's nice as is or pre73 is not something different that pre73.
Mr Williams: Do you know why ? changing tl072 by OPA2134, LME49720NA, not in audio path, not audible change but I got little improvement (6dB's) also in the noise floor. Could be possible only with my unit, not in generic units, due a faulty tl072 or another component (power issue) ?. I did for experiment and now I keep OPA2134, but I am not expertise or electronic, only I am a noob and don't know exactly why drop bit noise.
I will check Renesas for curiosity
PS: (*1) Cheap if you got 2SC3329 really cheapest. Beware on ebay for exclusive prices or fakes. A pair of dollar for a 30 or 40 batch, to match and put the matches not only high HFE, match hfe. Hard to find but not impossible.
Any real measurements done of how much the noise flow is changed? I did a measurement on the grounding patch and found out that it gave a noise floor improvement of -1.5 dB at 80 dB gain and max output. Not much but every dB counts and its easy and cheap.
Beware about TL072 replacement, is controversial question. I got 6dB noise drop replacing, but TL072 is not in the signal path. I don't know why, only test this.
In conclusion: A pair of bucks, pair of hours, and VERY happy.
PS: I have a pending, due to lack of time, learn more and return to this: the TL072 effect, why ? and the most important question for me, that still the damn hum, which was hidden by the high noise.
Nice. But have you done any measurement of the actual gain? My measurement is that the noise flloor is at -132.96 dBu with the ground patch. But that measurement is done with calculation of exactly 80 dB gain. That pertty good. With an other 5 dB it is too good to be true. So it must be so that the gain is not 80 dB. When change active components I would say it necessary to do that measurement. Did you change the bias on the input transistor? The CE voltage is about 2V. The 2SC3329 should have about 5-6 V for its optimal operation.
But have you done any measurement of the actual gain?
No, I can't due my home studio treatment (slow like a cathedral) except a SM7 and a little rane, all my gear are in a storage box.
I did a -75/4pm gain because is the max gain that I use (SM7) for very low thin sources. I did not intend to do it "scientifically" getting different measurres, just try this change and take action quickly.
Yes, I was adjusted the bias but not with a scope. I did not have it when I made.
Now I maintaing 2 issues that need learning and work. Currently only have 2 issues , I will try when I have my space "clear".
- Hum. I did the proposed ground change but is not enough.
- Distortion peek at HF changing TL072 that others IC's need bypass caps to power rails. But... TL072 is not in signal path, seems absurd, but in my case is an improvement of 6dB's.
Disclaimer: I always say "My" pre73, because I think there is less noise pre73 or different sonic characteristics depending on the batch that you purchased.
For example, when I took stock BC transistors, I measures the HFE and were very very different: 100, 200, 400 only a pair with 600, not compensated. The same with the ESR (too high) of one 22uf capacitor that I have yet to change.
I'm not saying this is bad, but denotes a disparity in each unit according to the lottery of components.
Maybe that's why there are two user groups: pre my stock is perfect or good, the other group, a bit disappointed, that do not like or need to change it.
When I asked for months if it did not seem noisy, was a controversial thread. Surely you pre73 is better than mine and I just had bad luck.
I spent weeks reading about changes, but I did not understand, if I want a Neve then I will buy a Neve.
I do not want wanna-be-Neve, I only wanted to improve aspects "measurables" for my cheap mic-pre that I bought only for push a SM7. It's that simple.
The improvement proposed by the Mr. Williams I think is simply the best of all, the most effective.
When I get a chance I'll tear it open again and play with the input stage biasing a bit with some Audio Precision noise and THD sweeps to see if that can be improved a bit.
I'm also going to replace the 22 uf caps with 100 uf, I want that depth I'm not getting. I already redid all the psu caps and bypassing. I used large value Pansonic FM caps with small bypasses. Maybe that's why I'm not getting any of the noise you are. Decoupling the 072 meter opamp with a 10 ohm resistor to the + power pin followed by another large el cap may also stop it's interference.
Jim, perfect! and thanks don't lose the focus then, don't hurry.
Aclaron, Bace: For curiosity, I ordered few toshibas from Futurlec. I guess they arrive soon and then I may explain to you what they are selling.
Don't expect them too soon, I ordered mine on Nov 1st, got notification they were shipping 3 days later, still haven't received them, I suppose they'll probably show up this week.
Still haven't found jack about the GR vs BL thing, I looked at the only datasheet I could find (from a random datasheet website) and it doesn't say either, and it is a Toshiba datasheet, though it looks kinda old.