SOLVED: Fix Relay Chattering or Chirping in Dayton Audio SA1000

[achugh]

Hi,

I am new to the forum. I have two Dayton Audio SA1000 sub-woofer amplifiers and both of them developed the same problem in last 3 months. They have lasted 10 years so I cannot complain about their quality. They outperformed their build, considering they were in use most of the time.

Problem
When I turn on the amplifier, I hear my relay click 2-5 times a second.

What I have tried so far

• I took (removed from the circuit board) the relay out and tested with a DC alkaline battery and the relay keeps its contact and does not make any noise. I tested it with a multi-meter in continuity mode and I hear my meter beep properly. I also checked the relay specs and the relay impedance of the coils match the specification. I also purchased new exact same relay. I tried to connect the new relay on the board without soldering and it exhibits the same problem. This has concluded that the relay is NOT faulty.

What I need help with

• Understanding what is this circuit trying to do so that I can figure out which component could have gone bad so that I can replace it and fix my problem.

NOTE: This circuit is traced manually from a double sided board so I may have a mistake. If things do not make sense please let me know and I can double check my traced circuit. I have already corrected few mistakes on my own so I think at this time there are no mistakes left in the traced circuit.

What I understand at the moment; which needs correction and help

• The circuit gets step down 19VAC from the transformer.
• The voltage then goes through a standard diode D6 (IN4004) acting as a clipper or half-bridge rectifier which is essentially converting the 19VAC into 19V DC with ripple.
• Next we have an electrolytic decoupling capacitor C6 (220uF, 35V, 85C) which is trying to reduce the AC ripple going into the 12V DC voltage regulator U1 (L7812).
• The regulator output is connected to C7 (100uF, 16V 85C) electrolytic filtering capacitor.
• Next we have standard (4 color) R6 (1.5K) resistor in parallel which I think is here acting as a bleeder resistor (or capacitor discharge resistor).
• Next we have D4 Zener diode (I can not see the part number) which is connected to C9 (220uF, 25V 85C).
• Then I have Q2 (PNP Transistor A1015) connected to D3 Zener diode (I cannot see the part number) with some resistors and C8 (47uF 25V 85C) capacitor. I think this is the start of protection circuit. In case D6 rectifier diode burns out, we will get full AC signal coming out of L7812. My understanding is that if Q2 turns ON properly then it will feed into Q3.
• Q3 is a NPN Transistor (C1815) which is connected to R5 (10K) current limiting resistor. My understanding is that if D6 burns and Q3 turns on, it will prevent Q4 from turning on hence the relay will not turn on at all.
• I can confirm D6 is working as expected so I thinking Q4 (another NPN transistor C1815) is turning ON which then makes the 12V DC output to be sent to the relay.
• I think D5 (IN4004) connected very close to the relay is acting as the flyback diode to prevent the rest of the circuit from getting damaged when the relay is turned off and the EMF (Electro Magnetic Force) energy is converted into very high voltage.
• C11 (100pF Ceramic) capacitor is there to clean any high frequency noise due to transistor switching.

Because the relay keeps turning ON and OFF at a very high speed, I am assuming that the 12V DC is no longer smooth and the voltage ripple is going below the threshold voltage that makes the relay noise (Constant clicking or Buzzing Noise).

What I would like to learn from the experts here is:

• Is my understanding of the circuit correct?
• If yes, then I think either C6 or C7 or C10 could be the culprits as to me they are supposed to do the ripple filtering and keep the Q4 transistor ON. Am I correct?
• Are there any other components here that would cause a ripple on the output if they fail?

NOTE: For diagnosing I only have a multi-meter available. If I had an oscilloscope, I would be able to see and confirm a bit more of my understanding but I do not have access to it. Since this is a very simple circuit, I am not sure if someone here have the parts to test it out on a breadboard.

Thanks for your help in advance.

UPDATES
09-JAN-2022: Updated the post with D5 and D6 part numbers. They are IN4004.

SOLUTION
31-JAN-2022: Replacing the C6 capacitor solved the problem. See https://www.eevblog.com/forum/index.php?topic=306719.msg3977495#msg3977495

[abdulbadii]

Would please RE relay too ?   ZD4 seems bit weird

[CaptDon]

I don't think you have correctly drawn the circuit. I can't figure out if that is some weird voltage doubler or what from your drawing. Looks like something is wrong on the low side of the drawing.  Please RE relay??? WHAT??? Can't begin to understand that reply??

[Audiorepair]

R6 is possibly connected between regulator out and Cathode of D4, you have messed up your diagram somewhat.

Better try again.

[commongrounder]

I would strongly suspect C6. Half-wave rectification causes plenty of ripple current which will age an electrolytic cap relatively faster. Your multimeter can show you the ripple voltage on its AC range. If your getting several volts across the cap, then it’s done. Not sure about the rest of your diagram, but the load current is likely pretty low for the designer to deem half-wave rectification sufficient. Also, if any of the caps are near a heat sink, they’ve probably gotten baked to death, and should be replaced.

[achugh]

I don't think you have correctly drawn the circuit. I can't figure out if that is some weird voltage doubler or what from your drawing. Looks like something is wrong on the low side of the drawing.

I have double checked the connections again and I am still not finding any mistakes. I am also not able to figure out the low voltage side which is why I am looking for help.

R6 is possibly connected between regulator out and Cathode of D4, you have messed up your diagram somewhat.

Better try again.

I double checked it with multi-meter in beep mode. I hear a beep when I connect the lead at the bottom of R6 and the other lead to the L7812 back which is also connected to the 0V of the transformer line. So I can confirm that the connection for R6 is correct. It is NOT connected to Zener diodes.

I would strongly suspect C6. Half-wave rectification causes plenty of ripple current which will age an electrolytic cap relatively faster. Your multimeter can show you the ripple voltage on its AC range. If your getting several volts across the cap, then it’s done. Not sure about the rest of your diagram, but the load current is likely pretty low for the designer to deem half-wave rectification sufficient. Also, if any of the caps are near a heat sink, they’ve probably gotten baked to death, and should be replaced.

When I connect the multi-meter to C6, C8 and C9 (with all of them in-circuit) I see that my meter will start increasing the resistance till it reaches its limits of 2M Ohms. This makes me think that these 3 capacitors may be good.
When I connect the multi-meter to C7, I get about 1.3K resistance. For C7 test as the current will flow the path of least resistance, getting 1.3K resistance makes me believe that I am measuring really R6 as that is the exact value I am getting instead of my meter trying to measure C7 in-circuit.
When I connect the multi-meter to C10, I get about 0.9K resistance. For C10 test as the current will flow the path of least resistance, getting 0.9K resistance make me believe that I am measuring really the resistance between the BASE and EMITTER of Q4 instead of my meter trying to measure C10 in-circuit.

Basically none of the capacitors are recording a short. Also from the pictures you can see that visually there is no damage. So has one of transistors (Q2, Q3 or Q4) failed?

NOTE 1: The above readings were done when the circuit was not powered on.

NOTE 2: The below readings were done when the circuit is powered on using DC Voltmeter setting

• When I connect my meter positive lead to anode of D6 and negative to cathode of D6, I observe that my meter shows -20.6V. I am thinking it is because during positive cycle the diode shorts and sends the signal forward. During negative cycle the voltage goes via the connected multi-meter hence the meter is recording negative voltage. Is my understanding correct?
• When I connect my meter positive lead to the anode of C6 and negative to cathode of C6, I measure 19.6V going into L7812 voltage regulator.
• When I connect my meter positive lead to the anode of C7 and negative to cathode of C7, I measure 11.75V coming out of L7812 voltage regulator. This is also the voltage I am getting at the end of the circuit going into the relay which is supposed to get 12V DC. Because I cannot see the switching of the transistors or AC ripple passing through my relay is constantly turning on and off like crazy.
• When I connect my meter positive lead to the anode of D4 and negative to cathode of D4, I measure 0.17V across it.
• When I connect my meter positive lead to the cathode of D4 and negative to resistor R6 point connected to 0V line, I measure 11.38V across it.
• When I connect my meter positive lead to the cathode of D4 and negative to resistor R6 point connected to 0V line, I measure 11.38V across it.
• When I connect my meter positive lead to the COLLECTOR of Q2 and negative to EMITTER of Q2 point, I measure 11.04V across it.
• When I connect my meter positive lead to the BASE of Q2 and negative to COLLECTOR of Q2 point, I measure 10.92V across it.
• When I connect my meter positive lead to the anode of C8 and negative to cathode of C8, I measure 11.04V across it.
• When I connect my meter positive lead to the anode of D4 and negative to cathode of D4, I measure -0.1V (negative) across it.
• When I connect my meter positive lead to the BASE of Q4 and negative to EMITTER of Q4, I measure 0.7V across it.
• When I connect my meter positive lead to the positive rail of R5 and negative to the other side of R5 which is also connected to the BASE of Q4, I measure 11.04V across it.

Some additional observations

• The output of L7812 voltage regulator is 11.75V DC instead of 12V DC. Is that my hint that the C6 or C7 or both are not doing the filtering they are supposed to do? Am I supposed to get 12V DC?
• All components after C7 thru Q3 are some kind of protection circuit because I get 11.75 V across R5 + C10 where R5 measures 11.04V and C10 measures 0.7V.
• Q4 turns ON to provide the small current for the relay with 11.75V to turn it on. This is why this transistor exists so the inductance (coil) in the relay does not burn due to heat dissipated if the output of L7812 is directly connected to it.

I was able to read that D6 and D5 are both IN4004 diodes. I have updated the original post with this information. I still cannot read the Zener diodes. These are the only components I don't know the details at the moment.

This is as far as I can guess with this circuit and need further help.

[commongrounder]

As I suggested in my earlier post, using the *AC* mode of your multimeter, measure across C6. In AC mode, your meter should ignore the DC component of the voltage and just measure the AC component (ripple) across the cap. This will help you visualize (since you don’t have a scope) how well C6 is filtering the half-wave DC from the diode. The relay is chattering. This means the DC supplying it, or the circuit controlling it is unstable. 78xx regulators can get very unstable when fed poorly filtered DC. You are missing a lot of information by measuring just DC voltages and resistances. Measuring charge time with an ohmmeter is not a reliable indication of capacitor health. A bad cap may measure normal or even high in value, but have terrible ESR.  Look for the simple solutions first.

[CaptDon]

Look at your schematic. There is NO WAY to get current flow from the bottom side of that relay back to the negative rail of your power supply EXCEPT for the charging current going through that one capacitor and once it is charged the current flow will stop and it will be impossible for the relay to remain energized. Either you have the circuit drawn wrong or it is the biggest waste of parts to create a 'momentary pull-in' of the relay at power up. Is that supposed to be a speaker protect relay to prevent 'turn on thump'. If you drew it correctly it is the dumbest relay circuit I have ever seen.

[abdulbadii]

that's why

Would please RE relay too ?

[CaptDon]

Looking again, even the charging current of C9 won't pull the relay. All of those parts that are multiples of 1000's of ohms can't provide enough current to operate even the most feeble of mechanical relays. Something is way off here, Schematic? Component Values you have indicated? It can't work as drawn.

[commongrounder]

This circuitry can only begin to make sense to me if the emitters of Q3 and Q4, and caps 8 and 10 negatives are connected to the power supply return rail. Then I can see the zener/PNP cap circuit acting as a timer that energizes the Q3,4 relay circuit after C9 charges up.

[achugh]

Thank you to all of you who are helping me troubleshoot this circuit. I truly appreciate your extra pair of eyes.

As I suggested in my earlier post, using the *AC* mode of your multimeter, measure across C6. In AC mode, your meter should ignore the DC component of the voltage and just measure the AC component (ripple) across the cap. This will help you visualize (since you don’t have a scope) how well C6 is filtering the half-wave DC from the diode. The relay is chattering. This means the DC supplying it, or the circuit controlling it is unstable. 78xx regulators can get very unstable when fed poorly filtered DC. You are missing a lot of information by measuring just DC voltages and resistances. Measuring charge time with an ohmmeter is not a reliable indication of capacitor health. A bad cap may measure normal or even high in value, but have terrible ESR.  Look for the simple solutions first.

I will put the whole thing back together and get the AC measurements soon.

Look at your schematic. There is NO WAY to get current flow from the bottom side of that relay back to the negative rail of your power supply EXCEPT for the charging current going through that one capacitor and once it is charged the current flow will stop and it will be impossible for the relay to remain energized. Either you have the circuit drawn wrong or it is the biggest waste of parts to create a 'momentary pull-in' of the relay at power up. Is that supposed to be a speaker protect relay to prevent 'turn on thump'. If you drew it correctly it is the dumbest relay circuit I have ever seen.

I believe you are talking about C10 here. C10 is rated for 50V, since the output from L7812 is 12V DC, C10 will never be fully charged. Will that allow the current to flow because it is NOT full?

Looking again, even the charging current of C9 won't pull the relay. All of those parts that are multiples of 1000's of ohms can't provide enough current to operate even the most feeble of mechanical relays. Something is way off here, Schematic? Component Values you have indicated? It can't work as drawn.

In my circuit diagram, I have not mentioned the relay which I understand is cause of confusion. The relay used by this circuit is SANYOU SLA-S-112DM. As you can see from the datasheet of this relay it only requires 75mA of current and 9V to 12V DC to turn on. Once it is turned on, it allows 30A of current on its load side. I believe this is why the output of L7812 is not being feed into this relay as it will seriously deteriorate the life of the coil in the relay with all the power and heat dissipation and instead this transistor circuit is used to drive this relay.

This circuitry can only begin to make sense to me if the emitters of Q3 and Q4, and caps 8 and 10 negatives are connected to the power supply return rail. Then I can see the zener/PNP cap circuit acting as a timer that energizes the Q3,4 relay circuit after C9 charges up.

I have checked it again, the drawing is correct. I plan to upload pictures of both side of the board along with Photoshop traces to highlight the traces of all of you can also see the physical circuit along with the easy drawing of schematic for understanding.

[coromonadalix]

I would still question the drawn schematic,  q4  emitter still need to be grounded,  it can not work  passing thru  d3 and c8,  you have 2 main relay coil pcb trace


For a start  i would simply change the electrolytics  and try,  it is an  rc timed  relay,  nothing fancy

If it still doesn't work,  i would test in diode mode all the transistors,  giving you did not make any probe slip while testing

[CaptDon]

WOW Achugh, you really don't have any clue of electronics or the circuit you are troubleshooting!!!! Without the emitters of Q3 and Q4 being returned to your '0v' in your schematic the relay can't possibly pull in. 12vdc applied to your 9-12vdc relay will not hurt it REGUARDLESS of what the contact rating is. YOUR SCHEMATIC IS WRONG as drawn. It is electrically impossible for that relay to have enough current sourcing to pull in EVER. And C10, it could be rated at 1 million volts, it doesn't matter since it will never see more than the .7vdc difference between the emitter and base of Q4. Why did they even use a 50vdc part??? 10vdc would have been overkill.
 

[abdulbadii]

easier to look if Q's output common grounded

[CaptDon]

Abdul's diagram makes more sense. In the end that whole relay circuit is only there to keep the speaker from getting whacked by the turn on transients. Looks like a delayed turn-on relay whose contacts are probably directly in line with the speaker voice coil. Horrendous turn-on thumps are characteristic of many poorly designed power amps and delayed turn-on relays (and often 'quick' turn off also) are very common even in reputable good quality amplifiers. In good amplifiers the relay is often de-energized by detected faults like D.C. on the speaker output or shorted condition overloads on the output. So now back to the question at hand, as others have said, replace the capacitors and it will almost surely start to work correctly again. All of my Carver Clair CBA-1000 amplifiers have one hell of a whack at turn-off. All of my speakers are more than capable of handling the thump without damage. It is annoying and there were some factory mods to reduce the thump. It was never worth the time to do the mods however. Any decent front-of-house speaker network could easily handle the thump.

[CaptDon]

R2/Q2 area still not drawn correct, C8 area not correct either, but getting closer to what would be expected.

[commongrounder]

Abdulbadii’s circuit is getting closer, but R1 is missing, along with its connections.  I believe another part of this circuit is to cause a quick collapse of the relay drive when AC power is removed. This opens the speaker relay to cut off any amplifier turn-off thump. It would be instructive if the OP would do a resistance measurement between the emitter of Q4 and the “0v” point. I’d be willing to wager there is a direct zero-ohms connection there.

[abdulbadii]

Sorry, clearly a Kicad noob..

[achugh]

Sorry, clearly a Kicad noob..

Thank you abdulbadii. I can confirm this trace is good. As others have pointed out, I made a mistake in tracing the circuit.

Abdul's diagram makes more sense. In the end that whole relay circuit is only there to keep the speaker from getting whacked by the turn on transients. Looks like a delayed turn-on relay whose contacts are probably directly in line with the speaker voice coil. Horrendous turn-on thumps are characteristic of many poorly designed power amps and delayed turn-on relays (and often 'quick' turn off also) are very common even in reputable good quality amplifiers. In good amplifiers the relay is often de-energized by detected faults like D.C. on the speaker output or shorted condition overloads on the output. So now back to the question at hand, as others have said, replace the capacitors and it will almost surely start to work correctly again. All of my Carver Clair CBA-1000 amplifiers have one hell of a whack at turn-off. All of my speakers are more than capable of handling the thump without damage. It is annoying and there were some factory mods to reduce the thump. It was never worth the time to do the mods however. Any decent front-of-house speaker network could easily handle the thump.

You are correct, there is a thump protection that I remember. This thing has been down for almost 9 months that I do not clearly remember all the conditions for thump protection. In some cases, the thump will not happen while in other conditions it does happen. The relay directly feeds the speaker with all the power for sure.

So is the consensus to order all the capacitors and start replacing them one by one till this thing starts working. Should I also order C11 ceramic capacitor or only electrolytic capacitors? Any reasons to doubt the BJTs?

When ordering new capacitors, other than the ratings (say for C6 220uF at 35V and 85C or higher tolerance) and size do I need to pay attention to anything else like Ripple Current, Impedance etc. I am looking at Digikey for placing order and chose Rubycon caps.

C6 Rubycon 220uF, 35V 105C - 35ZLH220MEFCT78X11.5
C7 Rubycon 100uF, 16V 105C - 16YXF100MEFCT16.3X11
C8 Rubycon 47uF, 25V 105C - 25YXF47MEFCT15X11
C9 Rubycon 220uF, 25V 105C - 25ZL220MEFCT78X11.5
C10 Rubycon 10uF, 50V 105C - 50YXM10MEFRTA5X11

In case I should be ordering C11 ceramic cap as well, I do not understand I can chose one. I see 101 written on the light brown disc and nothing else. So I selected the following TDK one as I could not find a Rubycon

C11 TDK 100pF, 2KV - CC45SL3DD101JYVNA

Now that we have this correctly traced circuit, can someone here explain what is this circuit doing as in how it is supposed to work? I would also love to learn what is happening.

Thank you for continuing to help me understand and fix my problem.

Updated on 30-Jan-2022 Corrected the temperature ratings of all capacitors to 105C instead of 85C which was a typo. Original caps were 85C, all the new caps are 105C as you can see from the datasheet links.

[commongrounder]

No need to replace the ceramic cap. They age well. I would recommend the 105 degree versions of the caps. Those have a much longer lifetime, especially for the half-wave filter. Replace them all. Do it one at a time if you want to pin down which cap exactly was the problem, but swap them all out. They are cheap.
Analyzing the circuit operation precisely would be easier if we could determine the zener voltages.

[cnqhdszq]

Please  upload the correct schematic ，thanks a lot !

[cnqhdszq]

I have not the software  ,so ,I can not open the  TXT files.

[achugh]

No need to replace the ceramic cap. They age well. I would recommend the 105 degree versions of the caps. Those have a much longer lifetime, especially for the half-wave filter. Replace them all. Do it one at a time if you want to pin down which cap exactly was the problem, but swap them all out. They are cheap.
Analyzing the circuit operation precisely would be easier if we could determine the zener voltages.

This was a TYPO. I corrected the post with an update message that all new caps are 105C.

I replaced the very first C6 caps which is acting as the filter for the half wave rectifier and the amp started working without any issue.

I have TWO (2) amps with exact same problem. So now I have a working amp and a bad one. Is there anything I can measure in circuit say AC or DC voltages to figure out the values of Zener Diodes. I do not want to remove them from the circuit as I do not have a variable DC power supply to properly measure them even if I remove it.

I am hoping that if we measure the voltages across a working and a non-working amp maybe someone here can make a guess what this circuit is really trying to do. I would still like to learn even though I now know the fix for the problem.

SPECIAL THANKS to commongrounder and abdulbadii for all your help and support as well as all the people who chimed in to help me get this problem fixed.

SOLUTION FOR RELAY RINGING or CHATTERING
Replace your C6 cap (this is 220uF with 35V) and that solved the problem for me.

Please  upload the correct schematic ，thanks a lot !

I have not the software  ,so ,I can not open the  TXT files.

You do not need to open the text file. I am able to see the diagram in this post. The TXT file for the software was provided in case anyone wants to edit the file further. There is no need to edit this file as the circuit is now correct.

[tronboy_rf1]

Hello.

I am WAY late to this party, but I was looking for some information related to this same amplifier (different topic) and was compelled to post some additional information here for the benefit of the community.  I have two of these SA1000 amplifier also - purchased 6-7 years apart.  The first one began to fail to enable upon a trigger input, and it was 4 days before the national championship football game (my alma mater was playing).  Since there was no way I would fix the original amp before the game, I purchased a second amplifier and shelved the first amp to be repaired later (which is now).   I replaced all of the electrolytic caps on all 4 boards in the amp (except for the low-loss electrolytics, which are orange in both of my units).   As part of this exercise, I noticed that the neutral lead from the mains plug was switched (front panel switch) instead of the hot side (which leaves the unit hot if there was ever some sloppy hands poking around that happy to close a neutral connection around the switch).  I thought this might be a factory error on the first unit, so I opened the second unit and found it was wired the same way.   So I left it.  The upside is that the hot lead does not run through the switch on the front panel.

All of this is backstory to what I wanted to point out for this thread.  I continued to poke around in the replacement unit, and I noticed a few capacitor "upgrades"  as compared to my original unit.  Most were just higher temperature capacitors and a vendor swap to a Chinese vendor that is actually still in business.  But the most notable upgrade was the 220uF/35V capacitor in the half wave rectifier ahead of the regulator for the delayed turn-on of the protection relay.  In the new unit, the value had been changed to 1000uF/35V, which obviously improves the ripple voltage since this is only half-wave rectified.   I made the same upgrade in my recap. 

In addition to my trigger enable problem, I had several bulged capacitors near the linear regulators on the preamp board (with the RCA connectors and switches).    I tested value, ESR/diss factor, and leakage for everything I took off.  I also tested several of the low-leakage capacitors and they were all fine (expected).    I had bad and questionable capacitors all over the place (like 470uF capacitors testing at 10uF, etc.). 

Good luck.