I have removed the dirt and I may try some experiments tomorrow where I try ensure the inrush current is limited so I do not make more damage to relay contacts. It may need an exchange anyway.
I did no fault as I can see. I waited to test until the PSU's were fully installed to prevent any errors like if I tested them "on table" using "loose" wires etc.
I have contacted them for a solution.
I have contacted them for a solution.
I got a response from Connex. It was to check a resistor and an E-cap which makes the delay. RC time constant I assume without knowing how it is connected. But I could measure that the E-cap was connected parallel to the relay coil. I was not able to trace how the smd resistor is connected other than it is 100 ohm as written on it. I would expect it to be connected to the E-cap. I could not trace it on the good one either.
I did a test more using a vario transformer that turns on during 0 VAC to make startup softer than just random on mains. This functioned perfect. SMPS started up as expected without a "big bang". Then I tried again just via mains contact where turn-on is random. Then it also worked as expected. No big bang and I tried several times.
I have two explanations for this:
1: The cleaning of backside PCB "did something" (shown in a picture above). But how this could influence a RC-circuit I don't know.
2: When I tested first time I just plugged in the mains connector to the mains in the wall. This could have "fooled" the delay circuit. On the other hand it worked perfect with other SMPS where I did it several times to test startup (the outlet I used is one that is always switched on and was closets to where I sit (lazy)).
But.....so far so good.......but I wonder how relay contacts looks after the "big bangs"......
Apart from this I like the idea to bypass/short the thermistor after it has done its job instead that it is always in the line and reduces the mains a bit and burn some heat. It also raises the impedance a bit. Another advantage is that it can be retrigged as it cools down immediately. Then it will work if a short cut-off in mains occurs.
Just in case I ordered a few of the relay used.......a little more than 1 USD each......
I did a test more using a vario transformer that turns on during 0 VAC to make startup softer than just random on mains. This functioned perfect. SMPS started up as expected without a "big bang". Then I tried again just via mains contact where turn-on is random. Then it also worked as expected. No big bang and I tried several times.
I have two explanations for this:
1: The cleaning of backside PCB "did something" (shown in a picture above). But how this could influence a RC-circuit I don't know.
2: When I tested first time I just plugged in the mains connector to the mains in the wall. This could have "fooled" the delay circuit. On the other hand it worked perfect with other SMPS where I did it several times to test startup (the outlet I used is one that is always switched on and was closets to where I sit (lazy)).
But.....so far so good.......but I wonder how relay contacts looks after the "big bangs"......
Apart from this I like the idea to bypass/short the thermistor after it has done its job instead that it is always in the line and reduces the mains a bit and burn some heat. It also raises the impedance a bit. Another advantage is that it can be retrigged as it cools down immediately. Then it will work if a short cut-off in mains occurs.
Just in case I ordered a few of the relay used.......a little more than 1 USD each......
I got some more detailed explanation from Connex regarding "soft start" of the SMPS.
For now I just continue my built and rely on that SMPS's works as expected.
"The soft-start limits the inrush current to about 20-30A for the first half cycle, depending on the temperature of the thermistor when is fist powered on and the voltage on the bus capacitors. It is very normal for a SMPS of this power to have the inrush current about 3-7 times the maximum operating current. Without an inrush current limit the current will be in the order of 150-250A which can blow the fuse and the rectifier bridge if the mains fuse doesn't trip first. Lower inrush current value isn't necessary and it can affect the reliable operation if the SMPS starts with maximum load, because the input current in this case also needs to add the current demanded by the load. For example if the load is 1000W with PF and efficiency taken into account the peak current must be >7.5A plus the current needed to fully charge the capacitors before the load draws current. With less than 10-15A inrush current it won't start.
The relay is supplied from the main transformer after the transformer starts working, and the transformer only works after the bus voltage has reached at least 70-75% of the normal operating value, at which point the soft-start already did its job and the relay can close.
The flash is more likely to have come from the fuse connector if the fuse wasn't sitting properly, because it is only kept by the spring contacts in place. Another unlikely cause can be that the thermistor failed if moisture was ingressed because when it powers on in the first 200-300mS there is a temperature spike of about 150-180*C which can evaporate the moisture and cause it to crack. It is explained in some application notes about this phenomenon.
The relay is supplied from that 100R resistor from the output of the aux voltage regulator and before the other circuits, isolated by D10 and D18 to avoid leaking the start-up current into the relay circuit and prevent the start-up so you can only measure the continuity to the 1500uF capacitor. The relay is connected in parallel with the 1500uF capacitor, which charges slowly and allows the ~400-600mS delay after power on before the relay closes.
If there was a short between dirty pins of the regulator it would kill the voltage altogether and the SMPS wouldn't start, so I guess it was just some flux residue.
The relay has 5 pins, two for coil and 3 for contacts, I doubt that is damaged, as I never had any relay fail in such an application and the peak current is carried by the thermistor, when the relay closes the current is already below its rated current."
For now I just continue my built and rely on that SMPS's works as expected.
"The soft-start limits the inrush current to about 20-30A for the first half cycle, depending on the temperature of the thermistor when is fist powered on and the voltage on the bus capacitors. It is very normal for a SMPS of this power to have the inrush current about 3-7 times the maximum operating current. Without an inrush current limit the current will be in the order of 150-250A which can blow the fuse and the rectifier bridge if the mains fuse doesn't trip first. Lower inrush current value isn't necessary and it can affect the reliable operation if the SMPS starts with maximum load, because the input current in this case also needs to add the current demanded by the load. For example if the load is 1000W with PF and efficiency taken into account the peak current must be >7.5A plus the current needed to fully charge the capacitors before the load draws current. With less than 10-15A inrush current it won't start.
The relay is supplied from the main transformer after the transformer starts working, and the transformer only works after the bus voltage has reached at least 70-75% of the normal operating value, at which point the soft-start already did its job and the relay can close.
The flash is more likely to have come from the fuse connector if the fuse wasn't sitting properly, because it is only kept by the spring contacts in place. Another unlikely cause can be that the thermistor failed if moisture was ingressed because when it powers on in the first 200-300mS there is a temperature spike of about 150-180*C which can evaporate the moisture and cause it to crack. It is explained in some application notes about this phenomenon.
The relay is supplied from that 100R resistor from the output of the aux voltage regulator and before the other circuits, isolated by D10 and D18 to avoid leaking the start-up current into the relay circuit and prevent the start-up so you can only measure the continuity to the 1500uF capacitor. The relay is connected in parallel with the 1500uF capacitor, which charges slowly and allows the ~400-600mS delay after power on before the relay closes.
If there was a short between dirty pins of the regulator it would kill the voltage altogether and the SMPS wouldn't start, so I guess it was just some flux residue.
The relay has 5 pins, two for coil and 3 for contacts, I doubt that is damaged, as I never had any relay fail in such an application and the peak current is carried by the thermistor, when the relay closes the current is already below its rated current."
When it comes to power-up test for the first time what is the lowest voltage I can use just to test if everything is connected correct and works.
Will do that before wires are connected to RCA and speaker binding posts. I will use a lab-supply for the first test as I by doing this I have a bit more control as I can raise the voltage slowly and use the current limiter.
Is -30V to low for initial test?
I have 2 x 0-30V PSU which can be paralleled. So -40, -50, -60 is also a possibility.
Will do that before wires are connected to RCA and speaker binding posts. I will use a lab-supply for the first test as I by doing this I have a bit more control as I can raise the voltage slowly and use the current limiter.
Is -30V to low for initial test?
I have 2 x 0-30V PSU which can be paralleled. So -40, -50, -60 is also a possibility.
connect your DVMs to proper points ( observing Iq and output node potential) and start gradually from 0V, rising up slowly to full rail voltage
same as that you have linear supply, and doing it with variac
and of course that you can daisy-chain ( in series) your two sections of lab supply, to get 60V
same as that you have linear supply, and doing it with variac
and of course that you can daisy-chain ( in series) your two sections of lab supply, to get 60V
Ok, will do that!
Regarding Mains Earth, Chassis, PSU Gnd I have wired Mains Earth directly to a screw at chassis which also goes to the earth connection on SMPS (all yellow/green wires). Then Drain on SIT goes to PSU Gnd (which is V+ on SMPS). It is my plan to continue the wire from V+ to go via a 10 ohm NTC thermistor to the Earth screw. Anything against this plan? .......this is what I have done on other amps and seems to work well.
Regarding Mains Earth, Chassis, PSU Gnd I have wired Mains Earth directly to a screw at chassis which also goes to the earth connection on SMPS (all yellow/green wires). Then Drain on SIT goes to PSU Gnd (which is V+ on SMPS). It is my plan to continue the wire from V+ to go via a 10 ohm NTC thermistor to the Earth screw. Anything against this plan? .......this is what I have done on other amps and seems to work well.
switcher PSU - are both rail ends )positive and negative) floating ?
I mean - I hope negative isn't connected to switcher case, which must be connected to safety GND
remember , if you use switcher PSU, then its positive rail is practically grounded, and its negative rail is practically negative rail for Lazy Singing Bush
I mean - I hope negative isn't connected to switcher case, which must be connected to safety GND
remember , if you use switcher PSU, then its positive rail is practically grounded, and its negative rail is practically negative rail for Lazy Singing Bush
Yes, I asked this question to Connex and V+ and V- are "nearly" floating. But there is a small cap and a resistor to the SMPS heatsink (which is now directly connected to chassis/Earth) to prevent static charge for components mounted on SMPS heatsink. They responded like this (what they call Gnd is V-):
"The GND is not directly connected to the heatsink but it has one 100nF capacitor and one 2.7k resistor to keep the voltage difference near zero and avoid building static charges. I can remove them then either the +Vout or -V out can be connected to the heatsink but the max voltage difference between -Vout and heatsink shouldn't exceed 100V."
I did not ask to get the 100nF and 2.7k removed as they have a protecting function. So I have from V- to chassis this 100nF + 2.7k. I don't think they will harm or make ground loop problem. I also measured that there are no DC connection from V+ or V- to heatsink, chassis etc.
The printing on the two output connectors are marked V- and V+.
"The GND is not directly connected to the heatsink but it has one 100nF capacitor and one 2.7k resistor to keep the voltage difference near zero and avoid building static charges. I can remove them then either the +Vout or -V out can be connected to the heatsink but the max voltage difference between -Vout and heatsink shouldn't exceed 100V."
I did not ask to get the 100nF and 2.7k removed as they have a protecting function. So I have from V- to chassis this 100nF + 2.7k. I don't think they will harm or make ground loop problem. I also measured that there are no DC connection from V+ or V- to heatsink, chassis etc.
The printing on the two output connectors are marked V- and V+.
A small correction. The 100nf and 2.7k is in parallel. So there will be these two from V- to Chassis. I have also found them on SMPS board. Two small SMDs.
Probably I should remove the 2.7k as it will get 65V across.....which is 1.5W.....is a lot for such a small device?
It will probably burn away anyway?
Now when SMPS is mounted it has done its job regarding static protection.......
They both sit so they are easy to remove.
It will probably burn away anyway?
Now when SMPS is mounted it has done its job regarding static protection.......
They both sit so they are easy to remove.
reconnect 100nF//2K7 from positive rail to switcher case
all good for safety, and then negative rail is fully floating
if more convenient, you can do that outboard ........ but remember you'll have that anyway, if you plan to use customary NTC between audio GND and chassis
all good for safety, and then negative rail is fully floating
if more convenient, you can do that outboard ........ but remember you'll have that anyway, if you plan to use customary NTC between audio GND and chassis
Yes, I have done the V+ to NTC connection. So 10 ohm when cold between V+ and Earth.
The tiny tiny SMDs are gone. I never found the 100nF. They are 1 x 1.5mm or so......so very tiny.
The tiny tiny SMDs are gone. I never found the 100nF. They are 1 x 1.5mm or so......so very tiny.
I wonder what the torque should be for the two screws which secures the THF51 via screw isolator and isolator pad. The force should not break the isolated or the isolator pad so I have tightened with "feeling". The split washer is pressed all the way and then a bit more.......
Then I check with DMM if it is isolated. I usually use the highest ohm setting like 20M as it will reveal the slightest conduction of current.
Then I check with DMM if it is isolated. I usually use the highest ohm setting like 20M as it will reveal the slightest conduction of current.
Well i would fully go by common sense, pure feeling, and by eyeballing only:
The beautiful Tokin shall be firmly seated without crushing the heat conducting isolator like som sort of a tire-tortured roadkill on the highway with all the guts all over the road/heatsink 🙂
I am going to lap the underside of my Tokins super flat with very fine grit ”sand paper” when the time comes for my build 🎷🙂🎸
The beautiful Tokin shall be firmly seated without crushing the heat conducting isolator like som sort of a tire-tortured roadkill on the highway with all the guts all over the road/heatsink 🙂
I am going to lap the underside of my Tokins super flat with very fine grit ”sand paper” when the time comes for my build 🎷🙂🎸
Ok, I think the pad should give a nice contact to THF51 and I think it was pretty flat from factory 🙂
I am close to make the first DC functional test using lab supply for one mono block.
If you lap the Tokin be sure to protect it against static charges......
From VFET amp I remember that the pressure on Sony VFET was less than expected as I exchanged one of the devices.
I am close to make the first DC functional test using lab supply for one mono block.
If you lap the Tokin be sure to protect it against static charges......
From VFET amp I remember that the pressure on Sony VFET was less than expected as I exchanged one of the devices.
I have got the heat transfer/isolation pads from Watanabe who sold them Tokins. Do you have these pads and? And are we also supposed to use cooling paste with them?