NTP M-100 Woes

[sardonic_z]

I'm working on an NTP EQ with those old opamps. A few of them were bad and I decided to make a few replacements. Wish I hadn't done that..

I'm using the m-100B schematics (available from the Gyraf.dk site). No matter what I do I can't get the BF115 and the BC307B directly after it to bias right. It's stable and not burning up, but there's no headroom whatsoever. It's definitely not a transistor orientation or a layout problem (I have checked and re-checked this about 100 times).

Anybody done this, or maybe have any ideas as to what might be going wrong? My suspicion is that somehow I'm using the wrong transistors or diodes. I was careful to match the values for modern silicon, but, still no joy.

 

[PRR]

> the m-100B schematics (available from the Gyraf.dk site).

That schematic has an error.

Shunt two Si diodes across the 3,6K resistor (much like the M100 below that). This will work. It is not fussy about transistors.

The M100B plan wants a second resistor C-B on the bias transistor. It is possible that the 3,6K should be installed C-B instead of C-E. But in real life, two Si diodes work very well.

 

[sardonic_z]

thanks, Santa! I'll give it a shot in a couple of days. Much appreciated.

 

[sardonic_z]

ok, so I tried the diodes and it does work, but the EQ still has no headroom. The only substantive difference between mine and the ones I'm working on is that I've used an MPSA18-D instead of the BF115. The hFE figures for these are way higher.. is it possible that it's clipping the next stage? Call me crazy but I thought that was a feedback network between B and C of that guy. Any help could potentially save my bacon.

 

[sardonic_z]

Hey there all, this is a bump on a thread that seems to have gone dead. Here we are, four months later, and the damned things are still in my life. As I mentioned before, I tried the aforementioned trick with the two-diode shunt with no success. I also tried a BF-115, for the hell of it (had to buy them on eBay from a guy in Greece). As before, all orientations have been checked and double-checked against the schematic, and my circuit design is duplicated more or less exactly from a working home-brew version of the opamp that was in place originally. I'm completely at a loss here. Is it possible that there is a badly-oriented transistor on the schematic? Could it be something else?

the schematics I'm working from:    http://www.gyraf.dk/schematics/NTP_M100_Schematics.pdf
the Gyraf Audio re-worked schematics:  http://www.gyraf.dk/schematics/NTP_M100_OPAMP.GIF

The only difference I haven't accounted for is the RC compensation between pins 9 and 11 on Gyraf's schematic. My understanding is that this is external, and probably has been implemented in the EQ itself, yes? I'd check myself but there aren't really any schemos floating around, and the unit itself is laid out so densely, and with such fragile PCB's that I've been avoiding pulling them apart other than swapping opamps. They both work when fully loaded with good opamps.. I have checked this several times.

Someone help, please!!

z

 

[etheory]

I've been working on an NTP M100 replacement also, and I currently have it at the proto stage, literally ready to solder together. My simulations show no signs at all of the issues you are speaking of.
My schematic is more or less the BC177B and 2N1711 / 2N2905 / BF115 Gyraf version (I think you'd call that the "A" revision, though I am aware you are speaking of the "B" version).

I managed to simulate the "B" version with the mod that PRR suggests successfully.
The two Si diodes across the 3.6K resistor seem to sort out the issues.
It appears that you might even be able to drop the BC307B + 4.7K resistor there entirely for those two diodes, though I'll sim to confirm this. This would place it topologically very close to the other version.

I will let you know how I get on and any info I can help with.

When you say "lack of headroom" can you qualify that any more? You say the original units work with the original opamps, when you say work do they show the same "headroom" issues?

Since you are working from a published schematic, would you mind sharing your layout and some photos so we might better be able to assist you fix your issue?

cheers,

L

 

[sardonic_z]

Here are the images. It's a clamshell layout, with the parts facing inward. On Side 1, the pins 1-6 appear R-->L as shown here, on side 2, pins 7-11 also read R-->L. The four empty pads at the top of each image are the connection points between the two sides. I'd like to point out that this layout was derived from a working model of the opamp, and then checked against the schematic.





The red part silhouettes are basically accurate, with the caveat that the flats are generally on the opposite side of where they are on the parts I'm working with (unless I'm dealing with a problem in my datasheets, which I very much hope is not the case). The BC237 is replaced with a BC550; all the other silicon is stock.

By "lack of headroom" I mean that the EQ I'm building these for distorts at pretty much all signal levels when I put them in place, as if the resting voltage was very close to the rail-to-rail limit. I've examined a recorded version of the sine wave I was sending through it and it appears squared off.

If you do see anything, by all means please let me know. I've got the guy who owns the EQ asking about it, and I need to give him an answer sooner than later.

thanks much,

z

 

[sardonic_z]

also: ignore the notes in the images. The details have changed since I made those.

 

[etheory]

If I get a chance tonight I'll do a sanity check for you.
FYI simulations indicate that for the M100B, removing the 3.6K, 4.7K and transistor between the output transistor bases and replacing that whole shebang with two Si junctions not only works, but improves performance and lowers distortion. Whether that translates in real-life, who knows.
L

 

[etheory]

I started tracing this out, then realized very quickly that at the rate I was going I'd be there all night.
To save me some time, how did you create this board layout?

Did you run a board check of the board after verifying the schematic?

From what I did trace you chose the BC307B splitting the 4.7K and 3.6K instead of the configuration on the schematic on Gyraf's website for sitting between the bases of the output transistors.
I feel that's probably the right way to go, and as I mentioned before, that configuration simulates fine.

So, pending more data, maybe a screen grab of the software you are using, I'm sorry, I gave it a good go, but it's going to be too time consuming to do a trace for you.

 

[etheory]

Hmmm, something is definitely fishy here.
I just hooked up my NTP M100 (A) variant on a breadboard, and it doesn't work.
Seems you have had more luck than that, but mine has issues that someone else on this forum had with the output pegging at one supply rail.

I soldered up, for comparison, one of my API 2510 variants, and it worked, on the same breadboard configuration, perfectly.
So there is something up with the schematic as I've interpreted it.

I have a feeling once I get to the bottom of it, it might solve your issue too.

I'll let you know.

 

[rafafredd]

Strangely, I once made a pair of mic preamps using the Gyraf NTP circuit and it worked very well.