Rebuilding a 2720 & Spice
Page 1 of 1

Author:  Michael-L [ Fri Dec 28, 2012 1:54 pm ]
Post subject:  Rebuilding a 2720 & Spice


I recently dragged my 2720 out of the closet and I'm in the process of rebuilding it. Lot's of flashbacks to '75 when a group of college students built it from the Radio-Electronics article. We bought the keyboard, case and controller board, but the rest we built from scratch, even etched our own boards.

OK, enough nostalgia.

The keyboard and controller are now working, the replacement 2N2222 and 2N2907 on the regulator are doing well.

I'm having problems with my 2 2760-2 VCO's. I've replaced some of the electrolytic capacitors and checked the power voltages (+/- 4.9V on th eboard), but I can't get them to oscillate at less than about 400hz.

Any thoughts on what is keeping the frequency so high? I've replaced C1 and cleaned the range and zero trim pots. I did have to use an LM741CV in place of the 748 but we also did that when we built it.

I've tried building an LTSpice model (its getting there, it actually oscillates now) to get a set of voltages to check against the boards to further trouble shoot. That's another story......

Thanks for any thoughts

Author:  PAiA-Scott [ Fri Dec 28, 2012 5:02 pm ]
Post subject:  Re: Rebuilding a 2720 & Spice

The 748 or 301 might make the difference. I think these have a pin 5 input that the 741s don't and this is the connection for the zero trim.

Author:  Michael-L [ Fri Dec 28, 2012 9:54 pm ]
Post subject:  Re: Rebuilding a 2720 & Spice


Thanks, my bad on the typo, it is the original 2720-2 VCO, not the A version. The R-E schematic doesn't show a connection to pin 5, and the zero trimmer is linked with the range trimmer to the -IN pin #3. The 2720-2A version of the has a trimmer on Pin #5 (Balance) and a 741 also has a Pin #5 (Offset Null).

I'll do some more testing with the 741 pulled out so that I can drive the voltage to the constant current source for the UJT directly to see if the limitation is the frequency of the relaxation oscillator or the 741 control of the the charging circuit.

Author:  PAiA-Scott [ Wed Jan 02, 2013 11:14 am ]
Post subject:  Re: Rebuilding a 2720 & Spice

I took a look at my assembly manual schematic and see corrections to the labels for the Q1 base resistors with the resistor to plus as R8 27K and the resistor to IC1 as R9 47K.

The printed circuit board graphics show R9 alongside Q1 and this is the resistor from the base to IC1. It should be 47K. R8 is shown between R9 and the edge of the board and this is the resistor to the positive supply voltage and should be 27K.

The assembly instruction steps for the installation of these resistors are correct -- a 27K is to +V and a 47K is to IC1pin6; however, if built from the assembly manual schematic, or, the one in the R-E article posted on the PAiA site and linked below, the resistors could be in each others places. ... rticle.pdf

Author:  Michael-L [ Wed Jan 02, 2013 11:00 pm ]
Post subject:  Re: Rebuilding a 2720 & Spice


Many thanks for the help, I appreciate your digging into the archives to find your old notes. Swapping the resistors dropped the minimum operating frequency to 280 Hz before the oscillator cuts off. I had to roll back a transistor substitution to get to that point, now I'll check out other items.

The LTSpice model now predicts a 130 Hz frequency at .625 volts with the trimmers set to the midpoint, so we're making progress. (I do always take a model's output with a grain of salt, but it has been fairly accurate so far)

I'll go back to driving the oscillator directly to see how low it will go before cut off to try to isolate the problem, but based on earlier measurements today, I think the 741 is driving the voltages correctly at pin 6 (i.e. pin 6 goes to zero with a 5V control voltage and 3.4 with a 0 control voltage).

Author:  Michael-L [ Sat Jan 05, 2013 12:34 am ]
Post subject:  Re: Rebuilding a 2720 & Spice


Additional information.

I went to swap the transistors, and on a whim checked the resistance. Big issues, the 4.7k was 5.21k, the 47k was 51.7k, and the 23k was 31.26k measured out of circuit, so I scoured the area to get some good resistors to make the tests. Long and short, with the correct values, the circuit shifted frequency down about 10Hz but still cut off at .9V control input. The resistor drifts were al in the same direction and within tolerance of each other so I wasn't too surprised.

The 741 is driving the constant current source correctly based on removing the chip and applying a voltage directly to pin 6 and testing the voltage at the base of Q1. The direct voltage produced the same proportionate Vb as the circuit with the chip in, so I believe it is OK. Also the chip operates over the entire control voltage range, so it isn't the source of the oscillator cutoff. .9 volts control produces an output voltage at Pin 6 of a Vb of 4.2v.

So I dug in deeper. What appears to be happening is that at oscillator cutoff, the Q1 Veb voltage is just below .59v I started paralleling resistors to adjust the R10 to raise the Ve (to raise the Veb), but it always returned to .59v with control voltage around between .8v and .9v at oscillator cutoff.

Digging some more over the weekend....

Author:  Michael-L [ Sun Jan 27, 2013 4:13 pm ]
Post subject:  Re: Rebuilding a 2720 & Spice - Sucess with details


Thanks for the help. Pointing me in the direction of swapping the R8 and R9 and some related changes did the trick and I have two 2620 VCO's tuned and singing.

Here's the rest of the story....
I had to replace the Q2 UJTs, both were noisy and had weird cut-off conditions. I used an NTE6410 as the replacement, but although they functioned, they have different characteristics and changed the frequency of the oscillator and the output voltage swing.

I also replaced Q1 with an NTE159. The NTE159 mirrors the original 2N5139 very well. One of he originals was dead, and the other was used to assess the NTE159 in a side by side comparison.

I also had a mismatch of 1-748 and 1-741 in the two VCOs. I did a very historically accurate rebuild, and went to Radio Shack and got the new TI LM741/SN and installed them. RS was the source of many of our parts when we first built the 2700, so it was fitting.

With those changes, the VCO oscillated, but wouldn't tune.

I relearned (learned?) some circuit theory and with the help of the LTSpice model, was able to reset the feedback resistance on the inverting summer that the 741 forms. The modified feedback involved simply changing R6 to 33k.

The VCOs still wouldn't tune, and I realized that I had neglected the Zero offset voltage as part of the inverting summer (I thought it was a biasing voltage divider....well it sort of is) Changing R5 to 47k brought the required Zero Offset voltage down, so that it was less than V-.

So it is on to tuning the keyboard and testing the rest of the modules.

Thanks again

Page 1 of 1 All times are UTC - 6 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
./cache/ is NOT writable.