Influence of output relay contacts to power amplifier distortion

[pma]

My goal is to show how output relay contacts may worsen THD of a high power amplifier. The amplifier under test is this one:

https://pmacura.cz/DIY_250W_4ohm_amplifier.html

and the output relay under test is Finder 40.31.9.024.0000 type with AgNi contacts. It has maximum contact current of 20A and is rated at 10A ac current with resistive load. The tested parameter was THD vs. output power at 1kHz into 4ohm and 2.2ohm resistive loads. Measuring bandwidth was 45kHz.

1. Load 4 ohm

Distortion is same up to 20W/4ohm. Then we can see and effect of nonlinear contact resistance, increasing THD (orange plot). Measurements were repeated with consistent results.


2. Load 2.2 ohm

This is a demanding load to any contacts in the signal path. We can see that up to 10W is distortion almost identical, above 10W we can see effect of the nonlinear contact resistance and increased distortion (violet plot).


Output mechanical relays are a two-bladed sword. They are easy to use, however their contacts are nonlinear and deteriorate with number of DC fault current breakdowns, due to arcing on contacts. I have not found yet a contact material that would not worsen amplifier parameters.

 

[Head_Unit]

however their contacts are nonlinear and deteriorate with number of DC fault current breakdowns

So if you keep using this amp, we might hypothesize the phenomena would worsen? How old IS the test amplifier anyway.
And, inquiring minds want to know: what is the "break-in" time?!?

Actually it seems there would be a "break-in" but like speakers very very gradual I should think, and likely not perceptible except by Bat-man maybe.

 

[SoundsGood64]

The case measured by OP probably is not audible. Unfortunately output speaker relays can add so much distortion that it becomes audible and a real problem that interferes with listening. I ended up replacing speaker relays in Denon AVR-3808ci and Denon AVR-2808ci after 10-15 years of use. There is a blog post (https://blog.debrakeleer.com/search/label/Audio) that describes the problem and how to fix it. The problem is decreased sound out of a speaker that sometimes goes away suddenly by turning up the volume.

Both AVRs use the speaker relay with part Fujitsu "FTR-F4AK024T" (the actual part in my AVRs, apparently different from the AVR-3808 in blog above and in Denon service manual). The relay spec sheet states material is silver plating AgSnO2 with a max resistance of 0.100 ohm. When a voltage is applied across pins 1 and 6, one relay part provides a signal path for one speaker through pins 2&3 and another speaker through pins 4&5. I'm using the pin numbers from the relay documentation; the pin numbers in the Denon service manual are different. Listening to speakers through my AVR-3808ci, there was an audible problem for Front L (RL101 in Denon manual), Front R (RL101), center (RL102), and Surround-A R (RL103). Surround Back speakers (RL105) and Surround-B speakers (RL107) sounded fine. The AVR-3808ci offers two different sets (A or B) for Surround speakers. Based on this pattern of audio problems, I think the previous owner of the AVR-3808ci (bought it used) might have had a 5.1 or 5.0 system and left the Surround Back and Surround-B speakers unused.

I removed the relays from AVR-3808ci. Applying 24V across pins 1&6, I then measured the resistance in ohms across pins 2&3 and pins 4&5. Here is what I found:

Relay pins2&3 pins4&5
RL101 0.6 ohms 93 ohms
RL102 0.5 ohms >3000 ohms
RL103 4.5 ohms 3.3 ohms
RL105 0.4 ohms 0.3 ohms
RL107 0.5 ohms 0.0 ohms

I managed to clean RL107 and got pins2&3 down to 0.0 ohms. The other relays I replaced with new relays.


For AVR-2808ci, there were audible problems for Front-A L (RL401), Front-B L (RL402), Front-B R (RL402). I didn't detect a problem with center, surround, or surround back speakers. Nevertheless, I removed all the relays and tested them by applying 24V across pins 1&6 and then measured resistance across pins 2&3 and pins 4&5:

Relay pins2&3 pins4&5
RL401 0.5 ohms 3.2 ohms
RL402 >3000 >6000 ohms
RL403 0.3 ohms 1.3 ohms
RL404 0.4 ohms 20-40 ohms
RL405 0.1 ohms 0.9 ohms

I replaced all the relays with new ones that each had 0.0 ohms at pins2&3 and pins4&5, except for one relay that had 0.1 ohm, so within spec.

Both AVRs sounded great after the relay changes.

Is there a way to prevent the deterioration of relays?

 

[NTK]

Is there a way to prevent the deterioration of relays?

Use MOSFET as switches like @tomchr (Tom Christiansen, Neurochrome) does with his Guardian-86 speaker protector.

Guardian-86

The Guardian-86 is a high-end solid state speaker protection circuit board module for use in audiophile audio amplifiers. The Guardian-86 uses MOSFETs instead of relays to protect your speakers resulting in better performance and better protection of your speakers.

neurochrome.com

 

[SoundsGood64]

Use MOSFET as switches like @tomchr (Tom Christiansen, Neurochrome) does with his Guardian-86 speaker protector.

Guardian-86

The Guardian-86 is a high-end solid state speaker protection circuit board module for use in audiophile audio amplifiers. The Guardian-86 uses MOSFETs instead of relays to protect your speakers resulting in better performance and better protection of your speakers.

neurochrome.com

View attachment 530603

Thanks NTK for pointing out that MOSFET can be used as a switch. That could be a fun project.

For conventional speaker relays, would any of these behaviors help reduce the wear and corrosion of the relay contacts:
if I don't have music playing when I turn on the AVR,
if I don't have music playing when switching a speaker set on
if I don't attach speakers to AVR when power is on
anything else?

 

[NTK]

Thanks NTK for pointing out that MOSFET can be used as a switch. That could be a fun project.

For conventional speaker relays, would any of these behaviors help reduce the wear and corrosion of the relay contacts:
if I don't have music playing when I turn on the AVR,
if I don't have music playing when switching a speaker set on
if I don't attach speakers to AVR when power is on
anything else?

Minimizing the chances of arc damages should help. I guess (but not sure) putting a MOV or TVS diode across the relay may help protecting the relay contacts from inductive flyback, but probably not useful against DC arcing. I am afraid much of this topic is beyond my area of expertise.

Elliott Sound Products seems to have some interesting projects.

Loudspeaker Protection and Muting

ESP Project Pages - Loudspeaker Protection and Muting. Protect you loudspeakers from amp DC faults

sound-au.com

Project 198

MOSFET AC solid state relay - Complete design, suitable for high voltage, high power, low distortion audio

sound-au.com

 

[solderdude]

In my repair days one of the most failing components (over time) is relay contacts, switch contacts and wipers of volume-, balance-, tone-controls.
I always loved repairing Revox stuff. You could simply take the components apart, clean contacts and it was good for another 10 years or so.

Relays are problematic devices, at least over time.
Distortion, crackling, intermittent operation (needing a 'bang' on the device or having to turn the volume up before it works) etc. are mostly caused by mechanical devices.

Relays usually work fine in the first period of their life and do not cause audible (but as shown can be measurable) effects. The problem lies in aging.

Is there a way to prevent the deterioration of relays?

The manufacturer of the device selecting the right relay may help here. Mostly manufacturers like to keep the production costs low and may select non-optimal components for that reason (cheap) that may not be optimal for the job at hand.

Using the stuff NTK mentioned helps but usually is not a 1:1 replacement and not many designs use those circuits from the start.

Replacing the relays with better suited ones (different brand or model) that is pin compatible or can easily be made pin compatible is an option.

 

[MAB]

In my repair days one of the most failing components (over time) is relay contacts, switch contacts and wipers of volume-, balance-, tone-controls.
I always loved repairing Revox stuff. You could simply take the components apart, clean contacts and it was good for another 10 years or so.

Relays are problematic devices, at least over time.
Distortion, crackling, intermittent operation (needing a 'bang' on the device or having to turn the volume up before it works) etc. are mostly caused by mechanical devices.

Relays usually work fine in the first period of their life and do not cause audible (but as shown can be measurable) effects. The problem lies in aging.


The manufacturer of the device selecting the right relay may help here. Mostly manufacturers like to keep the production costs low and may select non-optimal components for that reason (cheap) that may not be optimal for the job at hand.

Using the stuff NTK mentioned helps but usually is not a 1:1 replacement and not many designs use those circuits from the start.

Replacing the relays with better suited ones (different brand or model) that is pin compatible or can easily be made pin compatible is an option.

Agree. Relays degrade and need to be cleaned or replaced. And often get overlooked as the cause of issues. Here are the mechanical relays in an old preamp before and after cleaning:

 

[SoundsGood64]

Wow although better than before cleaning (post 8), the after cleaning picture still looks bad, I'm assuming the black is corrosion. Did the sound improve after the cleaning or do you have any measurements?

 

[MAB]

Wow although better than before cleaning (post 8), the after cleaning picture still looks bad, I'm assuming the black is corrosion. Did the sound improve after the cleaning or do you have any measurements?

Cleaning only does so much.
I replaced this one, but had to make an adapter since the model is discontinued and I couldn't find a part with a compatible pinout.
Yes, the relay performed intermittently, and had a measurable degradation in noise and distortion. While I couldn't hear the mains hum or the HD, but I could hear it when it misbehaved and partially cut out, mostly just after powering on. Which is why I ultimately replaced it.

Most of the old gear I test and restore suffers from relay issues, and volume/tone control issues. The exception is Bryston, which after years had no measurable or audible issues, and tested. I have some of these tests linked in my about page. I get lots of flak for not blanket-replacing capacitors, but for sure I almost always replace relays.

 

[wwenze]

How often does relay have to switch before distortion even appear? I looked at a random few and they're rated in at least the thousands if not 2 zeroes more. For a home that switches on-off a few times every day that will go into the thousands fast, but what about people who 24/7 their amps?

 

[amirm]

How often does relay have to switch before distortion even appear?

Depends on how much load there is on it when turned off.

 

[Sokel]

Reed relays are way nicer than normal ones as they are hermetically sealed and some of them use very high quality contacts.
Their higher cost just reflect their quality.

Edit:, an enormous plus, is that they are nearly silent, only a faint "ktinck" can be heard some cm away from them.

 

[SoundsGood64]

Depends on how much load there is on it when turned off.

I think this means: if you are playing loud music while you suddenly turn off your amp without first stopping the music, the relay will corrode faster. So a good practice to prevent relay corrosion is to stop playback on your streamer, disc, or whatever, and then turn off the amp. Is this correct?

 

[wwenze]

With AVRs one can theoretically design the controller to mute the output before switching the relay.

But why waste another source of planned failure.