I double checked for short with a flashlight before powering up. But I will do continuity checks with each of the next elements for each transistor just to be sure. The rails have voltage until the protection cuts it. The DC on the output is not the same every time. I took two videos, on the first dc is around 1V while on the second it gets around 2V. And when the power is closed it gets negative! Analog meter shows DC on left channel (2.5V scale), digital meter shows negative rail voltage.
Creek Destiny - DC at output # 1 - YouTube
Creek Destiny - DC at output # 2 - YouTube
Creek Destiny - DC at output # 1 - YouTube
Creek Destiny - DC at output # 2 - YouTube
Could the bias potentiometer be responsible?
The protection kicked in when i turned the pot to try adjusting the bias. On the first power up with the new transistors all leds stayed green. After increasing the pot the amp failed. How can I tell when the multiturn pot is at its minimum setting (without measuring i mean). Turning it to one direction some time i hear a tiny click sound. Then turning it many times to the other direction i also get a click. Is this normal on multiturns? It never ends turning, just clicks.
The protection kicked in when i turned the pot to try adjusting the bias. On the first power up with the new transistors all leds stayed green. After increasing the pot the amp failed. How can I tell when the multiturn pot is at its minimum setting (without measuring i mean). Turning it to one direction some time i hear a tiny click sound. Then turning it many times to the other direction i also get a click. Is this normal on multiturns? It never ends turning, just clicks.
The clicking is normal. A 10 turn pot has been used to enable critical adjustment, so turning it from one end to the other (powered) is the last thing you should think of doing. It's not clear whether you actually did this powered or not.
Hi Jomor, nice little videos of the problem.
Yes, miniature multiturn pots have a clutch to prevent damage. Just listen and gently rotate the 15 turns between the first click in either direction. Continuing to turn beyond the click makes no change to the setting but they won't cause any problem if they are roughly the same resistance value and wired the same way as the original parts.
Sure, you can't see where they are set and that's why I have returned to simple 1-turn cermet pots but they are not cool or finely adjustable which seldom matters really.
The "service mode" switch is a mode control for the microprocessor which indicates fault codes of your amplifier and you must have the service manual to understand and be able to install any software needed if you goof-up playing with it and maybe kill the remote control and protection circuit operation. Don't touch this unless you really know what you are doing or have a repair technician that knows the system. Otherwise you could have to bin the amplifier or build it into something else.
You haven't told us yet, where you bought those Mosfets you fitted. If they were fakes and not the genuine type required, your problems are just as Mooly warned - it won't work because the control threshold voltage doesn't match the amplifier, so they turn full on, red light, short etc. even with correct bias voltage and no signal.
I wouldn't place much importance on the negative voltage swing at this point. The important point is that your first replacement Mosfets were not so bad, as you say and you need to replace what you have with genuine parts if needed, as you concluded earlier.
Yes, miniature multiturn pots have a clutch to prevent damage. Just listen and gently rotate the 15 turns between the first click in either direction. Continuing to turn beyond the click makes no change to the setting but they won't cause any problem if they are roughly the same resistance value and wired the same way as the original parts.
Sure, you can't see where they are set and that's why I have returned to simple 1-turn cermet pots but they are not cool or finely adjustable which seldom matters really.
The "service mode" switch is a mode control for the microprocessor which indicates fault codes of your amplifier and you must have the service manual to understand and be able to install any software needed if you goof-up playing with it and maybe kill the remote control and protection circuit operation. Don't touch this unless you really know what you are doing or have a repair technician that knows the system. Otherwise you could have to bin the amplifier or build it into something else.
You haven't told us yet, where you bought those Mosfets you fitted. If they were fakes and not the genuine type required, your problems are just as Mooly warned - it won't work because the control threshold voltage doesn't match the amplifier, so they turn full on, red light, short etc. even with correct bias voltage and no signal.
I wouldn't place much importance on the negative voltage swing at this point. The important point is that your first replacement Mosfets were not so bad, as you say and you need to replace what you have with genuine parts if needed, as you concluded earlier.
This is difficult to try and diagnose at a distance. The fact that the offset is actually quite low doesn't suggest a "hard" a fault. It would be interesting to dangle a scope in there and check for oscillation but also we have to remember that the amp did work OK before. That points to either something physical having occurred (broken print) or the non functional FET's having caused damage somewhere in the driver stages.
Pulling the FET's out and powering up shouldn't damage anything but I don't think it will prove much either.
The click is just the clutch in the pot slipping at the end of travel.
The switch and LED will be used to test various functions but without a manual I wouldn't like to say.
This is what I would do and remember, this is working blind on this, so no guarantees
The bias pot you have turned. Its a very low value so measure the resistance of the good channels pot in circuit and set the faulty channel the same. Use the same meter for the adjustment and keep the polarity of the leads the same from channel to channel.
Do you see the two diodes, D6 and D7 near the opamp ? I would isolate them (remove one end) and see if the amp at least powers up and stays on.
At ALL TIMES do not connect a speaker and make sure you use a bulb tester.
Hi Ian, the stock mosfets are HUF76639P3. When they burned weeks ago I replaced them with HUF76639P3 bought from Aliexpress.com (chinese seller, probably counterfeits). They didn't give me same readings as the stock ones but they worked on the amplifier, meaning that I could listen music from both channels, but I could not adjust the bias at all. So I decidet to replace them with STP40NF10 bought from Mouser.com (thus genuine STs). The amp powered up whithout the protection kicking in. Before fitting them I had the pot turned closely to minimum. There was no voltage drop so i tried increasing the bias and while turning the pot the protection was enabled. Since then I get 1-2volts output dc. Replacing the STP40NF10 with the counterfeit HUF76639P3 still gives me dc on the output.
The reason I was experimenting with not genuine HUF76639P3 (or the supposed equivalent STP40NF10) is that HUF76639P3 are unfortunately out of production
I was thinking that since the DC is 1-2Volts perhaps it could be a leak from one mosfet. I will remove the board and check carefully the tracks and solder unused chinese HUF76639P3
The click is just the clutch in the pot slipping at the end of travel.
The switch and LED will be used to test various functions but without a manual I wouldn't like to say.
This is what I would do and remember, this is working blind on this, so no guarantees
The bias pot you have turned. Its a very low value so measure the resistance of the good channels pot in circuit and set the faulty channel the same. Use the same meter for the adjustment and keep the polarity of the leads the same from channel to channel.
Do you see the two diodes, D6 and D7 near the opamp ? I would isolate them (remove one end) and see if the amp at least powers up and stays on.
At ALL TIMES do not connect a speaker and make sure you use a bulb tester.
got it
edit: about the bulb tester, it seems I can't power up the amp with a bulb in series with the mains' live. When I plug the power cable and turn on the mains switch the amp enters the standby mode (this is what normally does). Then when I press the standby button to get out of standby the bulb flashes for a fraction of a second and the amp returns to standby mode. The bulb doesnt allow the initial current needed so I bypass it with a switch for a second until the amp gets out of standby. I tried 25watt,40watt and 75watt bulbs (in case the different resistance could allow the initial current the amp needs) with the same result. I measured the power consumption of the amp during the seconds before the protection kicks in and its 24 watts if this means something.
Last edited:
The pot is 50ohm and adjusted to 13.3ohms on the good channel. I had the pot turned anticlockwise many turns before fitting the STP40NF10. I just verified by measuring that turning anticlockwise increases the resistance value presented to the circuit, so it must was nearly 50ohms when i fitted the mosfets. But I m not sure if this corresponds to increased or decreased bias current. I may have done the opposite of what you suggested.
It sounds like the quiescent current is higher than a bulb tester can support.
I would imagine that the higher resistance VR1 is set too, the lower the current. So 50 ohms should be giving the lowest current possible. Also VR1 will affect the current.
You are going to have to get this in a "workable" state and that means being a bit adventurous with the fault-finding. If the amp is OK to power up fully then I would monitor the bias of the good channel (voltage across CN2) and then turn the bias pot for that channel (VR1) to give around 2mv across the test points. If it won't go that low then leaving it on its minimum, adjust VR2 carefully to see if that pulls it lower still.
Having done that, measure the voltage across R25 and set the bad channel to the same voltage. Turn the bias pot on the bad channel to the same end as gives the lowest bias current.
If you can do that then the amp should run OK with a bulb tester in place.
I would imagine that the higher resistance VR1 is set too, the lower the current. So 50 ohms should be giving the lowest current possible. Also VR1 will affect the current.
You are going to have to get this in a "workable" state and that means being a bit adventurous with the fault-finding. If the amp is OK to power up fully then I would monitor the bias of the good channel (voltage across CN2) and then turn the bias pot for that channel (VR1) to give around 2mv across the test points. If it won't go that low then leaving it on its minimum, adjust VR2 carefully to see if that pulls it lower still.
Having done that, measure the voltage across R25 and set the bad channel to the same voltage. Turn the bias pot on the bad channel to the same end as gives the lowest bias current.
If you can do that then the amp should run OK with a bulb tester in place.
I did set the pot at 13.3 ohms just like the good channel's. I checked very carefully all joints and cleaned very well some flux it had remained. I tried to power it up again and guess what.. it worked. It stayed about 15 seconds all leds green. My joy was that high that i forgot to put the switch which bypasses the bulb at off position. But this was too good to be true, suddenly the protection went on slight smoke came out of an element I didn't see clearly which, but it was from the good channel and the mains fuse was blown because I didn't turn the bulb's bypass switch off quickly. Now I measure continuity (about 25 ohms ) between gate and source of Q15 and Q18 of the (not any more "good") channel. I haven't touched anything from the good channel all this time. Quite disappointed. I m gonna try to remove all mosfets and turn it on to see what happens. If it stays on, I could start thinking of adding two independent modules for the final stage.. Edit: but there is not enough room there so this is out of the question too.
Last edited:
This is why you have to be so careful and methodical and follow a definite plan. So now the good channel is faulty too... that's not good is it. There is going to be more damage to the circuitry before the output devices and these components are going to be a lot more critical with regard to substitutes.
I just wonder if this is going to be fixable worked on like this tbh. Why should the good channel suddenly fail too ? That doesn't make sense.
I just wonder if this is going to be fixable worked on like this tbh. Why should the good channel suddenly fail too ? That doesn't make sense.
It doesn't make sense to me either. But I never touched anything on the good channel. I didn't have a reason to. I was very careful even having my hands washed and touching the board only by the edges. OK i did not get the bulb in line soon enough, but this has been done at least 10 times the last days (powering up without the bulb I mean) and the good channel did not have any issues neither of them. I m not that experienced but I believe that circuits do not fail just like that without a reason. Either the design is too sensitive and too easy to get unstable, or my device is a part of a problematic production batch (for example there is a similar case here and another referring to a too sensitive protection here ) . So either the protection is too sensitive, or problems occur often and the protection does often what it has to do! I am a person who is obsessed with his personal objects in a degree of sickness, I can get angry if someone touches my speakers, my car or my keys. So I am quite careful with my devices by nature, this one had a bad ..destiny i am afraid. So much money and time lost, I could have equal sq with 1/3 of the money. I am also persistent so I will continue my effort tomorrow with a clean mind, even if it is a shot in the dark
That's quite a tale of woe over on the Head-Fi forum.
Although I have no good reason for saying this, the circuitry does have that "finicky" look to it, and that it might not easily be reproducible.
It is up to you where you want to go next with it... my advice now would be to disable the protection and concentrate on just one channel.
Although I have no good reason for saying this, the circuitry does have that "finicky" look to it, and that it might not easily be reproducible.
It is up to you where you want to go next with it... my advice now would be to disable the protection and concentrate on just one channel.
(couldn't wait until tomorrow )
I removed all output mosfets. Two of the ex-"good" channel are gone. So I am only left with two genuine good ones. I tested all transistors and diodes around and all seem good, i get the same readings left and right channel. This is a good sign I suppose. I also tried to turn on the amp without the output mosfets fitted with no success. Either the microprocessor senses the mosfets absence or there is something else wrong too. I don't know what to to next except from ordering and trying IRL540N as Alex suggested and see what happens.
)I removed all output mosfets. Two of the ex-"good" channel are gone. So I am only left with two genuine good ones. I tested all transistors and diodes around and all seem good, i get the same readings left and right channel. This is a good sign I suppose. I also tried to turn on the amp without the output mosfets fitted with no success. Either the microprocessor senses the mosfets absence or there is something else wrong too. I don't know what to to next except from ordering and trying IRL540N as Alex suggested and see what happens.
Have you considered contacting Creek to see if they have any of the original output FETs available? Or if not, what do they suggest as a replacement?
Also, a vendor in the UK appears to have a few available. They are returnable if not original. Search item # 200950369260 on UK ebay. They do not appear to have very many available.
It might make the most sense to sub in the part Alex suggested and give it a go. At least you'd be dealing with original parts this way.
Lastly, you may need to add a Zobel network at the output to help calm the oscillation down.
Good luck!
Steve
Also, a vendor in the UK appears to have a few available. They are returnable if not original. Search item # 200950369260 on UK ebay. They do not appear to have very many available.
It might make the most sense to sub in the part Alex suggested and give it a go. At least you'd be dealing with original parts this way.
Lastly, you may need to add a Zobel network at the output to help calm the oscillation down.
Good luck!
Steve
Hi Steve, yes I have contacted Creek. It's a long story. I was in negotiation for sending the device to the UK but I recently asked for advice on replacement parts and I am waiting for an answer.
Thank you for searching on my behalf. I have already seen that and I was seriously considering buying them, but I was afraid that they might be bought from china and re-selled on ebay. But in the point I have reached, I will have to take further risks so I will contact them.
If they prove to be originals, I will have 8 pieces (including my 2 original mosfets) that I will have to preserve they won't get burned. Which means that I will have to solve the oscillation (or whatever it is) first with alternative mosfets. Difficult task if the alternatives demand modifications for tuning the bias.
I will need some help on that. I thought that the LR1 on the output was to control high frequencies.