I have built and am troubleshooting a QSXM2 preamp, using BC550 and 560 transistors in the front end for a MM cartridge. This is powered by a commercial +/-15v linear power supply. Both in an aluminum chassis with proper grounding.
I have two issues:
1. A large DC offset appears at output (both channels) of about +14V DC. This exists regardless of whether the DC servo op-amps are in or not. Adding a 10uf electrolytic cap in series with R93 eliminates this issue for now. Lifting R93 to separate the preamp and RIAA amp results in low DC offset on the main output. Without DC servo op-amps and the input shorted or not shorted, TP1 rests at about -20mv DC.
2. There is a 284MHz RF oscillation in the preamp section, both channels. Adding 0.1uf ceramic bypass caps as well as 470uf electrolytic caps at the V+ and V- input terminals does not quench the oscillation. Adding a 0.1uf ceramic bypass cap across the supply rails does not quench the oscillation. The only place on the PC board where adding an 0.1uf cap to ground to quench the oscillation is at TP1, but this will affect the frequency response in the audio band. There is minimal leakage of noise at the power supply terminals. C7, C17, C19, C21 are OK.
With headphones the preamp sounds good with the exception of added noise due to the oscillation. A sine wave oscillator at the input is reproduced without obvious distortion at the output, with the exception of added RF (a free transmitter!).
My main interest is the RF oscillation right now, then DC offset. Any suggestions are appreciated.
You seem to have the same trouble in both channels so concentrate to one channel to start with. Disconnect the power for the second channel.
First you must cure the very high offset.
Short the input to ground. Measure the voltages for the preamp. Start with this.
Greetings Per Anders!
Here is my QSXM2 preamp section voltage check report.
Power applied to left channel only
IC1 (OP07C) inserted. IC5 not inserted.
B&K model 391 DVM
V+ at X3: +15.002V
V- at X3: -14.993V
Actual voltage readings in (). Questionable results have "!" added
DZ1, 6,96V (6.870v)
DZ3, 7,08V (6.606V)
DZ4, 7,09V (6.617V)
R1, -9,5mV (00.00mv, input is shorted to ground)
R6, 491mV (492mv)
R7, 438mV (800mv)!
R8, 5,85V (5.633v)
R9, 5,88V (5.590v)
R10, 5,86V (5.551V)
R11, 5,90V (5.643V)
R12, 4,80V (4.575V)
R13, 4,83V (4.619V)
R14, 200mV (320mv)!
R15, 299mV (241mv)
R16, 5,64V (5.384V)
R17, 5,61V (5.391V)
R18, 5,64V (5.327V)
R19, 5,61V (5.360V)
R20, 5,64V (5.409V)
R21, 5,61V (5.441V)
R22, 5,64V (5.330V)
R23, 5,61V (5.274V)
R24, 508mV (254mv)!
R25, 518mV (262mv)!
R26, 1,65V (811mv)!
R27, 9,15V (14.058V)!
R28, 3,65V (203mv)!
R29, 9,12V (14.030V)!
R30, 3,62V (185mv)!
R31, 1,68V (823mv)!
R32, 1,56V (550mv)!
R33, 1,58V (540mv)!
R34, 926mV (44mv)!
R35, 926mV (53mv)!
R38, 834mV (117mv)!
R39, 855mV (71mv)!
R42, 325mV (430mv)
R43, 319mV (351mv)
R44, 321mV (447mv)
R45, 319mV (514mv)
R46, 316mV (561mv)!
R47, 323mV (500mv)
C3, 6,46V (6.100V)
Vbe, T1 580mV (632mv)
C4, 6,40V (5.970V)
Vbe, T2 550mV (553mv)
C19, 5,28V (6.453V)!
C20, 5,30V (6.440V)!
H1, 1,82V (what and where is H1?)
PREAMP_OUT_L, <+-0,2mV, changing with temperature fluctuations and air movements
(preamp side of R93 lifted -182mv)
DC-ADJUST_L, -5,66V, can go between +12-13V to -(12-13)V depending of balance. (fixed and steady -14.311V)
Extra DC Voltage readings to ground:
T1 BC560 collector 5.839v
T2 BC550 collector -5.803v
T13 BC550 emitter 3.822v
T14 BC560 emitter -3.674
T15 BC550 emitter 3.720v
T16 BC560 emitter -3.650v
Junction R50 and R51 -17mv
Try to figure out if the DC servo range is too small.
If the IC5 is not populated: Apply +15V at pin 5, check output and see if it moves
Apply -15V at pin 5, check output and see if it moves
I recommend that you use a variable power supply for this test so you can adjust the voltage in pin 5 from 0 to 15 V and 0 to -15 V.
Hi Per Anders!
Thank you for your suggestion and guidance.
With input shorted to ground:
Added variable 0 to +15v power supply to pin 5 of IC5 socket. No change at output at TP1 (which is at -184mv). IC5 socket (no chip inserted) pin 2 = 00.00mv. Pin 3 = 00.00mv.
IC5 socket (MAX420 inserted) pin 2 = 00.00mv. Pin 3 = 00.00mv. Pin 5 = -371mv. TP1 is at -183mv.
Where does input to IC5 come from? Where does IC5 output enter the preamp for DC correction?
Sorry, I was mistaken here. The opamp in mind here is IC1. Is it populated? If yes, please measure pin 2, 3, 4, 6 and 7?
More tests on IC5:
Add -15v to IC5 socket (no IC inserted) pin 5, TP1 moves from -180mv to -190mv
Add +15v to IC5 socket (no IC inserted) pin 5, TP1 no change T -238mv.
Shorting IC5 pin 5 to ground, TP1 = -237mv.
New tests on IC1:
IC1 (OP07C inserted), input shorted to ground, IC5 socket empty:
Pin 1 = +14.975
Pin 2 = -1mv
Pin 3 = -149mv
Pin 4 = -15.011v
Pin 5 = +0.2mv
Pin 6 = -14.321v
Pin 7 = +15.033v
Pin 8 = +14.919v
With IC1 removed, TP1 is at -300mv. Adding -20v to IC socket pin 6 (no IC inserted), TP1 reduces to -170mv.
Added 22K across R29, added 100k across R27. Bench supply of -9.5v at pin 6 now creates 0v at TP1!! With IC1 (OP07) installed, TP1 is now -7.5mv. Good! Will this increase noise from OP07 to preamp?
Inserting a 1KHz sine wave to the input causes IC1 servo to oscillate at about 40hz. Adding 10uf at the output of R29 eliminates this. Using an oscilloscope, I now see crossover distortion on the test sine wave at the card output, possibly from the RIAA stage-will look at this later.
If this servo change is approved by you, we now need hints on eliminating the 284MHz RF oscillation (this is not from a Radio Sweden International transmitter, but generated in the preamp). With the input shorted, the RF signal is 70mv p-p at TP1.
Good that you have found a point where the servo doesn't hit the roof. The output offset in TP1 should be the value which the OP07 has, single millivolts is fine. The DC servo doesn't contribute to any noise since it's heavily filtered. Nice that one channel works. Now you have to check the other channel.
Do you have the oscillation in both channels? Do you have an oscilloscope capable of 350 MHz or more? Make sure your ground wire of the probe is wounded around the probe and connected very near the TP1. A long ground wire serves as an antenna. If possible you should use a ground needle on the probe tip.
Hi Per Anders:
Thank you for quick replies today. This is enjoyable.
I edited my previous post as you posted, please read again.
Yes I have an older analog oscilloscope that is good for RF. The scope case is grounded to the preamp chassis case, and the scope probe ground lead is connected to the ground pin of J2. The preamp card Chassis ground connection is connected to the preamp chassis.
Both preamp channels oscillate. I checked with a spectrum analyzer too. Good transmitter at 284MHz!
If you turn off the power, do you still have the oscillation? For being an oscillation with BC transistors I think the frequency is very high. 1-10 MHz seems more natural.
The oscillation goes away when the power is turned off.
Definitely coming from the preamp sections, not the RIAA stages. I lifted R50 to separate the gain stage from the emitter followers. The oscillation appears at the emitter follower section. I confirmed that C103 and C104 are 100nF.
I also am a radio ham and broadcast engineer, and work on RF on both my radio hobby and at work, but there with slightly higher power of 88kW. :-)
Emitter followers can oscillate under certain conditions although it's much easier with power mosfets. How near the pcb have you mounted the transistors?
Each transistor is mounted between 0.6cm to 1cm above the board.
I thought being an emitter follower, which normally having gain less than 1, could not oscillate?
If you use the keywords "emitter follower oscillation", I'm sure you can find something.
I think you can put that aside and try to get the RIAA stage to get working properly.
Hi Per Anders:
Adding 330PF capacitors directly at the base of T35 and T36 (paralleling C17 andC18 except close to the transistors) reduces the oscillation at the output of the preamp section without affecting the frequency response at 20KHz. Earlier I tried an 0.1uf capacitor and 10KHz was attenuated by more than half, so that was no good.
I am wondering if the physical distance of C17 and C18 to the bases of the emitter followers allows the oscillation?
On to the RIAA section!
The RIAA stage came up fine for the left channel. On the right channel RIAA section I still have some offset on the output when the preamp has 10mv at TP1. For now I inserted a 10UF capacitor in series with R93, and the main output DC is now fine.
On the preamp DC servo amp IC1, when my test oscillator is set to 20hz, I see a sine wave on the output pin 6, as well as a little at C9. Would this AC audio correction signal "subtract" the from the main signal low frequency audio? Either way I added 10uF across C9 to make the correction voltage pure DC.
I'm listening to the preamp now-sounds clean. I will then look for the cause of the right channel DC offset and eliminate the audio coupling capacitor.
Good progress this weekend! Thank you again for your assistance!!
Hej Per Anders!
I would like to increase the gain about 6dB. Would I increase R51 to do this?
I am using a MM cartridge and the preamp is working pretty well (still have an issue with the right channel DC offset without input termination).