ours sound like the 2*5*2*0?

[balluda]

[quote author="Bauman"]Hey,

Hi All .I´m building my first 2520 opamp and have some questions:
Should I connect two diodes in series on D1 ? Do I solder r15 and c3 together on the upper side ? Should I use two resistors on r4+r5 to sum the value (90k)or should I use one resistor ? I´m using Fabios 2520BCr.1

Yes, 2 diodes as D1. No you should solder R14+C3 and then solder each side to the board. R4+R5 must be like 20k or 22k, and make R3 the same.
With 82k as shown it doesn´t work welll...

[/quote]
D1=OK, r15+c3(I solder them in series and then connect one leg from r14 and one from c3? Is it the same procedure with c2+r10?) ,r4+r5 (is there an updated partlist or where do I find the changes? I found some here but I´ve missed the 20-22k stuff,maybe I´ve missed more....

Thanks/Balluda

 

[udo]

Hey Fabio!
Very nice looking metal covers! :thumb:
Where did you get those? Are you planning to sell them?
Cheers, Udo

 

[Bauman]

Hi Udo,

The same guy that do the 1U rack cases to me do this little box...
If there´s some interest I could sell some, no problem. The good thing about them is that they are ceovered with Nickel and the solder! :twisted:

cheers!
Fabio

 

[udo]

Hey Fabio,
I might be interested ... depends on the price?!
Didn't realize that the 1U rack enclosure was completely finished. Could be interesting for me as well. Do you have final pics? I believe I only saw the prototype.
Cheers, Udo

 

[balluda]

Hi,again ,I´m just about to solder my 2520 RC v1 opamps made by Bauman,and I have a question I need some help with. I´m not the experienced type....OK here we go: Do I solder r15+c3 and c2+r10 in series or paralell? / Cheers Balluda

 

[pmroz]

.
Balluda... both in series. Solder one leg of the cap to one leg of the resistor, then mount the other two legs in the holes provided on the board.
Paul

 

[balluda]

Thanks a lot !! Have you used 22k on R4+R5 and R3 instead of 82k? / Balluda

 

[pmroz]

.
On Fabio's BC boards, I have 20K for both R3 and R4/R5.
You could also post Qs like this in the A*P*I thread, as that one seems more current (and Fabio might jump in).

 

[Mike Keith]

The 990 Op Amp is a very good amp, Thats what I use on my JM 130 mic pre but with the APIish output trans for a little warmer sound, also omitted the servo circuit. IM also a big fan of the Re 16 input trans, very high headroom.. no pad needed.

 

[deveng]

Diyr's,

I've designed, bult and tested a new discrete gain block of my own design. During my many months of simulations, calculations, breadboarding, testing, listening, trial and error efforts, etc, I've found something that I've been unable put my finger on. Prior to designing my own, I simulated the Melcor 1731, a few versions of 2*5*2*0's from this site and a 990 as a benchmark for the best. Both the 1731 and 2*5*2*0 simulation circuits show a bump in the frequency response at somwhere around 10-20Mhz which is espcially prominant at low gains. The Bode plots show the typcial 20dB/octave rolloff at the upper cutoff frequency set by the feedback cap. But, at a gain of 2 these both display this bump at reference level! Now the 990 exhibits a picture perfect response (of course, Deanne was a genius). When I designed my gain block I initially had this same bump! What I find interesting is my design is not a clone of either the 1731 or the 2*5*2*0 and in fact is quite different. So this became a challenging puzzle to solve. Typcial miller feedback did not subdue this bump. I was finally able to get rid of it by adding a single resistor between the 2nd stage transistors and final stage. This works on the 1731 as well. Bench tests comparing my op-amp with the 2*5*2*0 show a very slight 0.5dB drop in my opamp's response at 30kHz which I suspect is a result of the compensation to tame this bump. My next step will be to put the bump back in and do identical comparison tests. Since these opamps are almost always buried in circuit between transformers the bump would be tamed without effort, but I'm curious anyway as to if and how this will effect the sound!

Has anyone else seen this bump? Also, what are your views on this?

regards,
Jeff

 

[louder]

alo
at last i finnished my 2520.
they sound very good.
r3=22k
r4+r5=22k
they sound almost identical to the 3124+,than the melcor.
more latter.
best regards
pedro

 

[rafafredd]

I was finally able to get rid of it by adding a single resistor between the 2nd stage transistors and final stage

This is interesting. Can you tell us more?

 

[Samuel Groner]

Has anyone else seen this bump? Also, what are your views on this?

Around unity gain (typically between 10 and 20 MHz for unity gain stable discrete opamps) there is usually a quite impressive accumulation of poles in the response of a two-stage amplifier (BTW, why the hell don't we call these poles zeros? Very confusing...). It takes considerable care to compensate for these and the compensation is very likely to be partial only.

If we'd had an o/l response with two poles only (i.e. a dominant pole and a second "parasitic" one) and sufficient phase margin (IIRC more than 62°) the closed loop response is flat. As the first constraint is never fulfilled in real live (and realistic simulations) we get a bump in the response even with phase margins above 90°.

Having a balanced VAS and/or a three-stage architecture (as the 2520 IIRC) doesn't make compensation easier as it introduces even more poles. There are other topologies that are much easier to handle--i.e. folded cascode and one-stage current feedback opamps. These have other disadvantages, unfortunately.

For audio use it is IMO wise to limit the bandwith of each amplifing stage to about 200 kHz--usually done by adding a cap in parallel with the feedback resistor. If this is done, any reasonable stable amplifier shows flat response and overshoot free impulse response at noise gains greater than about 3 dB. As we hardly ever need noise gains below 3 dB this bump is usually of no concern.

Samuel

 

[deveng]

Thanks for the good explanation Samual. I knew it was a troublesome pole but my problem was (without doing un-ending math calcs) trying to figure a way to solve it! Since my overall topology was already established and generally working quite well, I was looking for a "siimple" fix without redesigning the entire amplifier. This turned out to be not that simple. I had narrowed it down to a particular stage but changes to the cap value would not work. So, I hate to admit it but with some trial and error playing around, it finally came down to adding the series resistance and it worked (with a change in the miller cap as well to fine tune).

Since this pole is way out, I suspected it wouldn't be an issue anyway but as a curious DIYr I was determined to fix it anyway!

rafafredd,
The resistor in my design is actually in series from the 2nd stage collector to the 3rd stage base. Since my design is different from the others, it won't help to solve the problem with the 1731 or 2*5*2*0 versions. Although, through experiementation you may be able to find a similiar solution as well. It can be done!

regards,
Jeff

 

[Samuel Groner]

I would be interested in a schemo of your amp, BTW!

Another trick against strange bumps in the o/l response above a few MHz is a zobel network at the output--10 nF plus 10 ohm does usually well.

Samuel

 

[JRE Productions]

Does anyone have a final parts list for this? I read through all the posts, and there were quite a few changes to Fabio's boards. I have just recieved a few boards from a fellow forum member and will giving the opamp a try. I do not have enough background yet to make my own substitution, so I am looking for a list of what parts to stuff this board with.

Thanks!

Joe

 

[JRE Productions]

Does anyone have a final parts list for this? I read through all the posts, and there were quite a few changes to Fabio's boards. I have just recieved a few boards from a fellow forum member and will giving the opamp a try. I do not have enough background yet to make my own substitution, so I am looking for a list of what parts to stuff this board with.

Thanks!

Joe

 

[JRE Productions]

ok, I can at least ask a few educated (or not) questions. I have three sets of boards. Two I have schematics for and one I do not. Two are Fabios. 2520 R1 and 2520 BC R1 The third I believe is a Melcor version with only 04/02 and a symbol that looks like "e"

So for Fabio's. All the resisters, caps and Diodes appear the same in both schematics. And both schematics look identical. I see two differences. The component silk screen is a little different and the Transistors used.

So the circuit is identical except one version uses

SKA 4693, TIS 97, TIS98, 2N5087, 2N3053, 2N4036
the other uses
BC550, BC 560, BD 139, BD140

So does one version sound different than the other? Is one set of transistors harder to get than the other? Can I build both pcb versions with the same components, so I have 4 identical opamps?

HELP!!!

Joe

PS can anyone identify the third opamp? Schematics?

 

[mick]

Here's a modded parts list I used for Fabio's BC boards, the red one's are the changes that
I picked up from the threads, as of yet my boards have not been tested so I can't say if this is all present and correct.

C1 1000p
C2 47p
C3 47p
C4 39p
C5 10p
D1a 1N4148
D1b 1N4148
D2 1N4148
D3 1N4148
Q1 BC550C
Q3 BC550C
Q4 BC550C
Q5 BC550C
Q6 BC560C
Q7 BD139
Q8 BD140
Q9 BC560C
QCS BC550C
R1 150K
R2 360R
R3 20K
R4/5 20K
R6 3K3
R7 3K9
R8 1K
R9 56R
R10 800R
R11 5R6
R12 5R6
R13 56R
R14 1K
R15 1K
R16 47K
RCS1 2K7
RCS2 25K

Ok so my question is, do all these changes go for the PP R268 boards ? and was there any other changes that were
tested and worked for the PP R268 boards ? I know there was talk about R18(RCS2) but I don't know if it was tried
and liked and what not. Also near the Q1,3,4 part on the board there's a line going from between C5 and R2 toward Q9
is that a link wire ? I would be most grateful for any help here as any talk of this board seems to get mixed in between
all the others and it's difficult to decipher.

Thanks Mick

 

[CJ]

Thanks for that!
I have a pair of those guys that need stuffing.
Is that the same circuit as Melcor 1731?