Need a pair of eyes on a spectra sonics 101 vero layout

[orangechili]

Ok,
I have a vero layout I did i would like someone to take a look at for me. I've been going cross eyed trying to find an error in it as i'm having an issue building it up.  The vero is based off the page 7 schematic of this manual http://www.technicalaudio.com/pdf/SpectraSonics/SpectraSonics_101_amplifier_manual_schem.pdf

I have read the 8 page spectrasonics 101 thread here already and that was very helpful in putting this together.Ground connections follow the way it is on the original boards. i know there were quite a few variations on schematics with resistors, caps and transistors but the general connections seem to all agree together from what i see…

Here's the problem, I built it up and it gets sound but there is no difference in sound when I turn the gain pot, the volume remains the same except there seems to be more of a hum on one side of the rotation of the pot. I have triple checked the connections around the pot, transistor pinouts, etc.. Kinda stuck here. Is the schematic wrong or likely my vero layout abilities? Any help is appreciated.

 

[rockinrob86]

I can take a look later!

Would detailed pictures of the original's help?  I have a pair, but would be interested in making another pair with your layout when you get it working....


Did you adjust the values of the resistors to match the transistor's bias needs?  How do you have the pot wired?

 

[orangechili]

Nice! Reference to an Original would definitely help, i have a feeling one connection is just off. It passes audio just very low in volume and gain pot basically does nothing. I have the vero layout perfectly sized for the smaller size veroboard tayda electronics sells so it's easily mountable and cutting of the vero isn't necessary. Tried to make things easy and clean.

 

[orangechili]

I was thinking of putting some trimpots on the values that have variance between the schematics. The problem I'm encountering seems different than a transistor bias issue to me. I have the pot wired up as stated in the vero with a 5.5k resistor hanging off the one side. Perhaps i'm misunderstanding the wiring there but I read the larger spectra sonics thread and seemed that was the way to hook it up. I tried a resistor for fixed gain(43k for 50db) also and results were poor, lower volume than with the gain pot.

I'm mainly wondering if I messed something up with the connections from the schematic to the vero layout? I checked it so many times i'm questioning my sanity

 

[Whoops]

Hello,
I know that this it not a straight answer to your problem, but why don't you just copy the original PCB traces layout and do a self etch version?

Actually if you are doing a Spectrasonics project from scratch you should do the 110A instead of the 101, it's an improvement over the 101 and has better headroom.

Doing a self etch is quite simple, probably less hassle than doing in Vero board.

You just have to print the traces to a board and etch it.

 

[Whoops]

.

 

[orangechili]

Thanks for the 110 recommendation! I've seen threads on it but haven't looked into it much.

I used to self etch years ago but the muriatic acid/peroxide mix became too much for me to deal with. That's why for simple circuits i usually just stick to veroboard(while still observing the best grounding I can). If its larger I usually just buy a pre-fabbed board from a supplier.

 

[Whoops]

Thanks for the 110 recommendation! I've seen threads on it but haven't looked into it much.

"The Model 110A Preamplifier Module (right) is an updated version of the Model 101, which implements a bi-polar design (+/- 24V Volt). It was introduced in 1971, and was used in late Spectra Sonics console design. It's improved headroom and noise specifications earned it an even more "clean" reputation than the original 101 preamp. Its increased maximum output of +24dBu makes it extremely versatile, especially in modern digital recording studios"

The 110A also has a higher input impedance 2.74K instead of the 620ohms of the 101

 

[digitana]

Hi ..I checked the layout ..couldn't find any issues. I'll look again. Meanwhile  I added your  component  labels to the schematic attached in case anyone else wants to check.  I assume you have checked for any faults on  your  populated veroboard?

I am getting a couple of 110's soon. Was thinking of  doing a similar perf/turret board layout  for them using  the DIYLC software I think you used ?

 

[orangechili]

Thank you for updating the schematic to correspond with the vero layout! I have checked the build extensively for faults numerous times. Not saying it's not possible there is one, but can't seem to find one yet. I read on a thread somewhere of someone excited to do their own vero for a 101 and then they went silent. I'm wondering if they encountered a similar issue? I have to wonder if there's a flaw floating around on the schematics for these?,,, I'm knee deep in other projects right now but will try to experiment more this week with transistor bias /trimpots and see if I make any progress.

 

[ruffrecords]

I think you need to connect input- and output- to 0V.

Cheers

Ian

 

[digitana]

yeah..I'm busy on other builds too...      There is also the schematic which I think I got form GDIY somewhere..which seems to be the oldest (maybe original ?) with penciled in values...see attached.

This  looks the same as  yours and also the same as the one in the 101 patent of 1965..but I noticed that  it has a different resistor value  (your R9)  of 4.99k controlling the current through the 3 diodes, instead of  3.01k.  R9  limits the  current through forward biased D1-D3 , so  the exact  bias voltage at the base of  Q2  should  depend to an extent on the value of R9.  You could try adding a 2k  trimmer in series with your  R9 and see if adjusting it helps.

From what I have read the biasing of the  transistors in 101 and 110  is  maybe  where the mojo  is 
Particularly so if you choose random  samples of a given transistors which will  likely have a range of  gains . This would explain why some people  who have compared their 101 cards  with various existing  schematics have found different values of some resistors..presumably the ones  involved in biasing.

Once I get my 110 cards working I might measure the gains of  the various transistors  and compare the R values to those in the 110 schematics I have seen.

 

[digitana]

I think you need to connect input- and output- to 0V.

Cheers

Ian

yes agreed..it  is mentioned on the vero layout notes of orangechilli

 

[orangechili]

Yes I kept those grounds separate just like they were on the the original boards and then all grounds tie to one star ground point in the chassis per the instructions in the other 101 thread.

 

[Whoops]

Some info for consideration:

At the time the 101 boards were being made in production, transistor beta's varied all over the map ... typically from 50 to 350 between two units out of the same lot.  At Spectra Sonics, they used to grade the transistors and match them to betas within 5 points.  This matching is important, because, except for the input transistor, they function in pairs.  If you were to match a plow horse and a race horse ... to draw a comparison, it would be a disaster.  Whereas, matching two similar horses (or transistor current gains) would be a good solution.

I would suggest that you use the exact same brands as used in the 101.  It would be ok to substitute a pn3566 for a 2n3566 ...that sort of substitution would be ok.

Input transistors were selected, in circuit, for low noise.  Chances are, that is less of a problem than it used to be.  However, it is something to consider.

The brand and type of diodes is critical.  It has a notch in its temperature curve which is used to advantage in the circuit.

Be sure to use solid tantalum capacitors, and polystyrene caps where they are used.  These are important ... because they have the necessary bandwidth. While we only use the amplifiers in the audio band, the circuitry itself is exceedingly wideband ... typically 3 to 10 mHz ... and it requires capacitors which have that bandwidth. This wide bandwidth is one of the things that contributes to the low noise floor, and the fast peak overload recovery of the amplifier.

 

[orangechili]

Ok, I had some time to mess around with this today and made quite a bit of progress. Thank you to all for the info and ideas! I love this forum

I put trimpots  on R9(3.01k) and R5(10.7k) on my vero layouts. Those are the 2 resistors I saw variance between all the schematics in my searches. Still had low output and gain control was still doing nothing, trimpots did seem to help bias what was there a bit, didn't seem like a big deal too much, you can tell when it was out of bias.… On a whim I started tapping all the components on the board with a chopstick and lo and behold D1 made a loud pop sound when i tapped it. It looked fine from above but on the other side of it I noticed a little bit of glass missing. I replaced it with another 1n461 i  had and now all appears to be working fine. The gain control functions like it should now and output seems about right for the circuit. There is hum and a little noise but I have the circuit in a test box with transformers real close to the power transformer and power board. The next step is to put it in a new case away from the PT. I would venture to say the vero layout is verified now.

Great advice on matching transistors for the pairs Whoops

I used NOS 2n3566 and 2n4249 transistors and 1n461 diodes I got from a surplus electronic store here locally. I could not source 2n4248 transistors so I did sub them with 2n4249's. I am going to experiment with some modern type transistors as I have sockets on the board and the pinouts are pretty standard.

 

[orangechili]

I did match the transistor pairs(Q2,Q3 and Q4,Q5) and noticed an improvement in sound. I also tried out some matched BC550B's and 2 BC560C's I had laying around, The performance was pretty much the same as the 2n3566 and 2n4249 transistors to my ears. Now to just get this circuit away from the power supply and in a new box…

 

[orangechili]

I adapted the 101 vero layout to 110a specs per the schematic Whoops posted earlier in the thread. It is very similar in topology as far many connections go but has a few changes(bipolar power supply, resistor changes, etc) so wasn't too hard to do. Hopefully some people will get some good use out of it

It is currently unverified so if someone wants to double check my work against the schematic that would be awesome

Here's a link to the vero i use for these as well
http://www.taydaelectronics.com/hardware/prototyping-boards/small-stripboard-94x53mm-copper.html

 

[digitana]

great! glad you have the 101 running..Also thanks whoops for that valuable info.  Was planning on checking the betas of  the
npn and pnp pairs in the original 110 boards I am getting soon..  Original 70's NOS Fairchild 2n3566's seem to be getting harder to find  at least  in UK/EU..so  finding reasonably matched pairs from  might be  a bit hit and miss unless you buy lots. Maybe later production  including PN3566  versions  is the way to go...

Also thanks for the 110 layout.. I'll check when I get time..

 

[orangechili]

FWIW the new spectrasonics 101 cards look like they use the pn3566  transistors instead of the older 2n3566 ones. They also appear to have mica instead of polystyrene caps.