The insidious nature of protection relay failures

[mondialfan]

I had a protection relay issue crop up with one of my own amps recently. This is my personal Nakamichi PA-5, which is a Series 1 version, built in 1985. I pulled the amp out of my audio room to do some quick testing in planning some upgrades and the tests showed the early stage failure of the protection relay. I tested this amp about 18 mos ago and it tested perfectly fine at that time. I have seen this type of failure before in other folks audio gear that I have repaired so I thought I'd share this info. Many times owners will believe that as long as their protection relays are passing sound that they're working fine, but this can be far from the case. In a lot of instances by the time you're hearing actual audio drop-outs due to the protection relay contacts failing it is actually in its final stages of failure. In fact its likely been causing high freq distortion in your music for many hundreds of hours. Thus the insidious nature of these failures. Although the distortion is quite measurable, it is only nominally noticable, but it is affecting playback. It may show up as a perception of harshness or excessive treble in the music. Once the music stops all is quiet and everything seems fine.

The audio distortion from these failures can many times start in the higher frequencies and as the relay contacts get worse the elevated distortion begins to work its way lower in the freq range. I've also seen where the distortion only shows up at elevated power levels and the lower power levels measure fine. In this amp the opposite is true and the amp is measuring worse at the lower power levels and gets moderately better at higher output power.

Here is a graph I took showing 4 traces. The comments section in the bottom of the graph tells you what the plot lines are for. As you can see in the RED R-Ch trace the 20kHz THD+N is substantially worse especially at the lower output power levels (500mW - 5W) than the L-Ch 20kHz trace. This is over 15dB of difference between channels in the distortion at its worst reading. However at 1kHz the issue is completely gone as the two channels are very similar to each other with distortion only being a few dB dissimilar.

 

[mondialfan]

After taking the last measurements that were shown I then used the handle of a screwdriver and lightly tapped on the back side of the board directly opposite of the protection relay in the bad channel. The attached graph is the subsequent test after tapping the relay. As you can see the distortion is now completely gone at 20kHz. Needless to say I have a set of new relays on order from Mouser to install along with the upgrades I have planned.

 

[northpaw]

Thanks for this information. I see these relays (RL101 and RL201) are specified as 10A 250VAC relays, but no other info. From the circuit they should be SPST. I looked over inside photos I took of my unit, but unfortunately I did not take any that reveal label info on the relays, so I am at a bit of a loss to know model number, pin spacing, coil ohms, etc. I would like to order spares before opening everything up (the PA-5 is my main listener), so I'd greatly appreciate any info you might be able to provide, perhaps just a relay model number that I can look up to match the specs.

 

[mondialfan]

The original relay in the Nakamichi amps is a Matsushita JR1A 24VDC, (1 Form A) relay rated for 10A/250VAC. I picked the Omron G5PZ-1A-E DC24. This is a high capacity relay rated for 20A so it should hold up quite well.

I also bought a 3rd G5PZ to replace the relay in the AC soft-start circuit. It uses the same JR1A 24VDC relay. I expect it is in worse shape than the protection relays because it sees a substantial load everytime it's activated.

 

[northpaw]

Thanks so much for this info.

 

[Ray Gianelli]

I had a protection relay issue crop up with one of my own amps recently. This is my personal Nakamichi PA-5, which is a Series 1 version, built in 1985. I pulled the amp out of my audio room to do some quick testing in planning some upgrades and the tests showed the early stage failure of the protection relay. I tested this amp about 18 mos ago and it tested perfectly fine at that time. I have seen this type of failure before in other folks audio gear that I have repaired so I thought I'd share this info. Many times owners will believe that as long as their protection relays are passing sound that they're working fine, but this can be far from the case. In a lot of instances by the time you're hearing actual audio drop-outs due to the protection relay contacts failing it is actually in its final stages of failure. In fact its likely been causing high freq distortion in your music for many hundreds of hours. Thus the insidious nature of these failures. Although the distortion is quite measurable, it is only nominally noticable, but it is affecting playback. It may show up as a perception of harshness or excessive treble in the music. Once the music stops all is quiet and everything seems fine.

The audio distortion from these failures can many times start in the higher frequencies and as the relay contacts get worse the elevated distortion begins to work its way lower in the freq range. I've also seen where the distortion only shows up at elevated power levels and the lower power levels measure fine. In this amp the opposite is true and the amp is measuring worse at the lower power levels and gets moderately better at higher output power.

Here is a graph I took showing 4 traces. The comments section in the bottom of the graph tells you what the plot lines are for. As you can see in the RED R-Ch trace the 20kHz THD+N is substantially worse especially at the lower output power levels (500mW - 5W) than the L-Ch 20kHz trace. This is over 15dB of difference between channels in the distortion at its worst reading. However at 1kHz the issue is completely gone as the two channels are very similar to each other with distortion only being a few dB dissimilar.

View attachment 2470895

That is VERY interesting. While it makes sense that this would be, it never would have crossed my mind. Those graphs really tell a story.

Thanks for posting!

 

[Binkman]

Needless to say I have a set of new relays on order from Mouser to install along with the upgrades I have planned.

Great info... do you suggest your findings are Indicating a failure looming soon?
We've seen countless relay replacement threads (wish I had a dime per) what's your take on early audible sign or symptoms, if any, as most wouldn't even think of it or had the equipment.
Subtle audio signs...

edit: example... IMO on my post #17 is severe.
https://audiokarma.org/forums/index.php?threads/technics-sa-5760-very-scary-static-bursts.986494/
Seen this on DIY and might be same OP?

 

[mondialfan]

Great info... do you suggest your findings are Indicating a failure looming soon?
We've seen countless relay replacement threads (wish I had a dime per) what's your take on early audible sign or symptoms, if any, as most wouldn't even think of it or had the equipment.
Subtle audio signs...

edit: example... IMO on my post #17 is severe.
https://audiokarma.org/forums/index.php?threads/technics-sa-5760-very-scary-static-bursts.986494/
Seen this on DIY and might be same OP?

IMO when the relay begins to alter the original signal then it has failed and needs to be replaced or have the contacts cleaned if possible. The big problem is that many times this failure is stealthy enough and begins slowly and gets worse with time so you can become accustomed to any change in the sound. Because most folks do not have the test gear to see this problem in its infancy its appropriate IMO to change the relay(s) when one would normally change the caps and perform a common refurbishment. 30+ years is a long time for a set of relay contacts to be responsible for passing signals and not altering those signals through increased resistance or capacitance across those contact points.

The thread you linked to seems more like a component failure or connection issue somewhere. The OP replaced the relay but he could jump across the relay contacts as a test to eliminate the relay as a possible problem if still unsure. I would of course use a set of el-cheapo speakers and have them fused for that test.

 

[avionic]

have the contacts cleaned if possible

Contacts are pitted.. Cleaning is a short term fix. Unless you burnish down the contacts to eliminate the pitting. By that time, there may be no contact surface left. Best option to replace..

 

[gadget73]

I expect it is in worse shape than the protection relays because it sees a substantial load everytime it's activated.

probably looks worse but I wouldn't be surprised if functionally its better off. Each time it makes and breaks that heavy load it will arc and burn off the oxide layer so the contact surface is probably better functionally up to the point that its just wasted.

 

[mondialfan]

I just got around to changing these relays out a few weeks ago and I thought I'd post an update. The relays I recommended (Omron G5PZ-1A-E DC24) are not the same pin out, however they can be easily modified to work. One side of the relay is the correct pinout, the opposite side needs to have an extension leg added. I included a picture of how I did this. I used a piece of 20ga bare solid core hook-up wire. Bend an "L" onto one end on the hookup wire and put this L under the pin to be extended, then bend the pin over to capture the hookup wire and then solder the pin and hookup wire together. Then bend the other end of the hook up wire the length needed to reach the hole in the board. I sat the old relay next to the new relay and used that to gauge where to bend the hookup wire to fashion the new pin.

The picture of the new relay in the board shows the new wire coming out from the end of the relay to the old pin hole. The nice thing is that these relays have small standoffs along the edges so the wire only very slightly raises the relay off the board once its bent flat against the bottom of the relay and its almost imperceptible that its sitting up and slightly tilted off the board. Due to the fact these boards are single sided and all of the electrical traces are on the bottom side of the board you can run this wire against the top of the board with no worries about shorting.

I also used the same relay for the soft-start circuit. These relay coils had a nearly identical DCR as the old ones and operate perfectly, plus they're a higher capacity at 20A.

 

[pw5599]

Good info here, thanks mondialfan. I've recently re-capped and tuned up several of my old amps and receivers and as part of that, replaced the relays. And on many I thought they were still good but based on your tests, they probably weren't so was a good idea to change them. I had one in a Nikko Alpha 230 that wasn't available in it's original pinout configuration and had to do a "deadbug" style with wires going from this side to that on all pins but works great. When replacing relays one has to find the correct coil resistance, or close, as well as contact amperage and pinout, if possible.