Hi,
I'm currently repairing a Yamaha YST-SW800 subwoofer and have found some of the Mos-Fets to be dead but am unable to locate the ones that i require.
This is what the service manual says and the parts that i need-
Q16- V5877200- FET- 2SK2378
Q18- V5877200- FET- 2SK2378
Q19- V5877000- FET- 2SJ406
Are there alternatives to these?
I'm currently repairing a Yamaha YST-SW800 subwoofer and have found some of the Mos-Fets to be dead but am unable to locate the ones that i require.
This is what the service manual says and the parts that i need-
Q16- V5877200- FET- 2SK2378
Q18- V5877200- FET- 2SK2378
Q19- V5877000- FET- 2SJ406
Are there alternatives to these?
Are you guys sure IRF 640/9640 work in these amps?
I just soldered in a parallel pair of IRF9640 in the same subwoofer and the amp now seems to operate fully in class-B I think.
Not switching. Of course something else may still be blown in it.
From the datasheets it can be seen that IRF has much higher VGS than 2SJ/2SK at least...
And yes all IRF640 are N-channel, but an IRF640N is a bit different from IRF640. That is what was meant earlier in the thread
I just soldered in a parallel pair of IRF9640 in the same subwoofer and the amp now seems to operate fully in class-B I think.
Not switching. Of course something else may still be blown in it.
From the datasheets it can be seen that IRF has much higher VGS than 2SJ/2SK at least...
And yes all IRF640 are N-channel, but an IRF640N is a bit different from IRF640. That is what was meant earlier in the thread
Hi all
I'm currently repairing a YST-SW800 and have run into the same issue(s) as above.
I was wondering if anyone found any suitable equivalents in the end for the switching stages, because the 2SJ406/2SK3278 are "low voltage drive" MOS-FETs, and I can't seem to find any that are similar i.e. VGS(th) = 1.5 > 2.5V
I tried IRF640/IRF9640 but ran into the same trouble as above with no switching occurring and am assuming the difference in threshold voltages is the cause (will do some more measurements to verify this)
Did anyone manage to get this amp running in the end?
Cheers
I'm currently repairing a YST-SW800 and have run into the same issue(s) as above.
I was wondering if anyone found any suitable equivalents in the end for the switching stages, because the 2SJ406/2SK3278 are "low voltage drive" MOS-FETs, and I can't seem to find any that are similar i.e. VGS(th) = 1.5 > 2.5V
I tried IRF640/IRF9640 but ran into the same trouble as above with no switching occurring and am assuming the difference in threshold voltages is the cause (will do some more measurements to verify this)
Did anyone manage to get this amp running in the end?
Cheers
I have checked the schematic against the description of the problems you report.
There are 2 cases to analyze:
- Q16-19: Class AB amplifier. Upon inspection of the schematic I can't find a problem with higher Vgs-th MOSFET. There is a bias adjustment. If there is no way to bias the amplifier out of class B: R14 can be increased for more range (increase in small steps, turn bias pot to minimum before starting each test, check with multimeter for no mistake before power up). If there is no switching or no output I think the problem is not the Vgs-th of the class AB FETs but some other part damaged.
- Q22-25: Buck regulator. Current hysteresis driven. Upon inspection of the schematic I can't find a problem with higher Vgs-th MOSFET. Gate drive is suboptimal. Vgs is also Vds when switch is ON. Higher Vgs-th just reduces efficiency. But it wasn't high-efficiency to start with. If there is no switching or no output I think the problem is not the Vgs-th of the switching FETs but some other part damaged.
I assume you have checked the PSU is working. Bear in mind the SMPS has no current limiting, so it is likely to fail in case of shorted power devices in amplifier section.
EDIT: Due to the current hysteresis design of the switching regulators, these won't start switching with no signal unless the output stage is biased into class AB, above a certain threshold of bias like 100mA (2*50mA). It may not switch at all at lower bias levels like 50mA (2*25mA). A signal and a load is required to see the regulators switching.
btw: For perfectionists: To search for MOSFET able to operate at Vgs=5V search for "logic level MOSFET". However, from the point of view of semiconductor usage/performance the circuit is not perfectionist to start with.
There are 2 cases to analyze:
- Q16-19: Class AB amplifier. Upon inspection of the schematic I can't find a problem with higher Vgs-th MOSFET. There is a bias adjustment. If there is no way to bias the amplifier out of class B: R14 can be increased for more range (increase in small steps, turn bias pot to minimum before starting each test, check with multimeter for no mistake before power up). If there is no switching or no output I think the problem is not the Vgs-th of the class AB FETs but some other part damaged.
- Q22-25: Buck regulator. Current hysteresis driven. Upon inspection of the schematic I can't find a problem with higher Vgs-th MOSFET. Gate drive is suboptimal. Vgs is also Vds when switch is ON. Higher Vgs-th just reduces efficiency. But it wasn't high-efficiency to start with. If there is no switching or no output I think the problem is not the Vgs-th of the switching FETs but some other part damaged.
I assume you have checked the PSU is working. Bear in mind the SMPS has no current limiting, so it is likely to fail in case of shorted power devices in amplifier section.
EDIT: Due to the current hysteresis design of the switching regulators, these won't start switching with no signal unless the output stage is biased into class AB, above a certain threshold of bias like 100mA (2*50mA). It may not switch at all at lower bias levels like 50mA (2*25mA). A signal and a load is required to see the regulators switching.
btw: For perfectionists: To search for MOSFET able to operate at Vgs=5V search for "logic level MOSFET". However, from the point of view of semiconductor usage/performance the circuit is not perfectionist to start with.
Last edited:
Hi Eva,
I've completed some more testing today and the switching stages seem to now be working once their is enough current as you say, as shown on the trace below. The upper and lower traces are the switching rails B+ and B- on the schematic (x1) and the centre trace is the output (x50)
The behaviour of the bias adjustment however doesn't seem to be matching what is written in the service manual. The voltage measured between TP1 and TP2 is as follows under these various conditions:
Load (Ohms) | Output (VRMS) | Max possible adjustment range TP1-TP2 (mVDC)
Open | 0 | 0 > 0
Open | 25 | 1 > 6
Open | 35 | 8 > 10
16 | 0 | 0 > 0
16 | 1 | 56 > 57
16 | 10 | 68 > 68
16 | 20 | 68.5 > 68.5
8 | 0 | 0 > 0
8 | 1 | 62.0 > 62.5
8 | 10 | 70 > 70
8 | 20 | 70 > 70
This seems a little odd... I'm concerned that the bias cannot be adjusted between 50mV and 250mV when there is no signal, as is obviously the usual procedure, but the amp seems otherwise to be working. I will continue testing but thought I would let you know how far I had got because you kindly took the time to look into the circuit. Thanks
I've completed some more testing today and the switching stages seem to now be working once their is enough current as you say, as shown on the trace below. The upper and lower traces are the switching rails B+ and B- on the schematic (x1) and the centre trace is the output (x50)
The behaviour of the bias adjustment however doesn't seem to be matching what is written in the service manual. The voltage measured between TP1 and TP2 is as follows under these various conditions:
Load (Ohms) | Output (VRMS) | Max possible adjustment range TP1-TP2 (mVDC)
Open | 0 | 0 > 0
Open | 25 | 1 > 6
Open | 35 | 8 > 10
16 | 0 | 0 > 0
16 | 1 | 56 > 57
16 | 10 | 68 > 68
16 | 20 | 68.5 > 68.5
8 | 0 | 0 > 0
8 | 1 | 62.0 > 62.5
8 | 10 | 70 > 70
8 | 20 | 70 > 70
This seems a little odd... I'm concerned that the bias cannot be adjusted between 50mV and 250mV when there is no signal, as is obviously the usual procedure, but the amp seems otherwise to be working. I will continue testing but thought I would let you know how far I had got because you kindly took the time to look into the circuit. Thanks
An externally hosted image should be here but it was not working when we last tested it.
Sorry I forgot to inform you all.
I tried the amp without load and maybe there was a bad solder joint also.
After resoldering in the IRF parts and testing with load my amp started working fine!
50mV to 250mV across R29,R32 is only 3mA to 17mA, divided by 2 output devices. This is a low bias design due to the LF-only duty.
This is intended to be adjusted without signal, without load, no need for the regulators to be switching. Measuring with load and signal will give a different reading on R29,R32 actually more related to the mechanism of the regulators.
My previous post also explains the problem with bias and higher Vgs-th MOSFET. However, in case of no responsiveness to bias adjustment check also Vbe multiplier transistor (Q39) and potentiometer (VR1).
Actually the waveforms look a bit strange.
- How were the rail waveforms measured? Are you sure it is 5V/div? Are you sure these are the modulated rail waveforms?
- The +/-50V output waveform looks somewhat chopped. Is it the signal generator? The preamplifier section? Or the amplifier section?
This is intended to be adjusted without signal, without load, no need for the regulators to be switching. Measuring with load and signal will give a different reading on R29,R32 actually more related to the mechanism of the regulators.
My previous post also explains the problem with bias and higher Vgs-th MOSFET. However, in case of no responsiveness to bias adjustment check also Vbe multiplier transistor (Q39) and potentiometer (VR1).
Actually the waveforms look a bit strange.
- How were the rail waveforms measured? Are you sure it is 5V/div? Are you sure these are the modulated rail waveforms?
- The +/-50V output waveform looks somewhat chopped. Is it the signal generator? The preamplifier section? Or the amplifier section?
Thanks Eva, yes my main concern is that there is hardly any adjustment possible - the voltage measured between TP1 and TP2 is 0mV when there is no signal. Therefore I checked again for defective parts, but all the components in the bias network are OK.
I will check these components again because the bias is non-responsive...
Yes, the B+ and B- are at 5V/div and the output is measured at the amplifier output at 1V/div x50 ... these were measured at the nodes indicated in the service manual. The signal generator is Audio Precision, but the scope is quite old, I think the output waveform looks choppy due to the low resolution of the scope, there is no significant measured distortion on the output (distortion being measured by the AP)
I'll persevere.... thanks again for your help,
Hmm...I seem to have forgotten which mosfets I replaced. Switching or linear ones. I definitely did not replace all 8, only 4.
And I did get the bias right without modification.
I did not even find a photo of it anymore, sorry.
On the amplifier board I think I only replaced MOSFETs.
I do not 100% remember.
On the PSU board, which I think usually blows if amplifier has shorted output devices, I think I replaced power transistors, some resistors and the small electrolytic caps.
Hey, I've got the same subwoofer that I'm trying to repair, any chance you remember which replacement transistors you used for the power supply?
Or if anyone else can recommend alternatives?
Originals are 2SC4140 & 2SC4163 which I can't find from trusted sources.