Certainly my experience of the Mouser project being delivered to the UK was very impressive.
Ordered on a Thursday and parcel tracking showed Texas, Tenessee, Charles-de-Gaulle (France), then Birmingham (UK) before being delivered to me 100 miles away just outside London on the Monday.
Incredible.
Ordered on a Thursday and parcel tracking showed Texas, Tenessee, Charles-de-Gaulle (France), then Birmingham (UK) before being delivered to me 100 miles away just outside London on the Monday.
Incredible.
I wouldn't be overly worried about supply sequencing for a single LM3886. For multiple in parallel, you do have to worry about the supply sequencing.
For a single LM3886, I'd put a sizeable supply cap after the switch - say 2200-4700 uF - to smoothe out any switch bounce. If you do that, I'm sure a regular DPDT switch will be just fine.
Tom
For a single LM3886, I'd put a sizeable supply cap after the switch - say 2200-4700 uF - to smoothe out any switch bounce. If you do that, I'm sure a regular DPDT switch will be just fine.
Tom
Last edited:
Thanks Tom,
I guess you would want a pretty substantial switch if there were significant additional capacitance (thinking of folk adding 10's of k
as they do) so I wouldn't underestimate that aspect for long term reliability. A bit of C should certainly help smooth out any millisecond differences in contact timing.
Interesting one I'd just be uneasy switching high DC current this way.
I guess you would want a pretty substantial switch if there were significant additional capacitance (thinking of folk adding 10's of k
as they do) so I wouldn't underestimate that aspect for long term reliability. A bit of C should certainly help smooth out any millisecond differences in contact timing.Interesting one
I'd just be uneasy switching high DC current this way.
Unless you switch the amp off with the music cranked to 11 and happen to hit the switch right at the signal peak, you won't be switching very high currents aside from the inrush current to charge the caps. You could choose to worry about arcing in the switch, but that's easily remedied by a snubber/spark-arrestor across the switch.
Given the use of relays in the automotive industry, the loads switched by these relays, and how reliable the relays are in that application (when was the last time you replaced a relay in a modern car?), I suspect the worry over relay/switch reliability is a bit towards the neurotic end of the spectrum. Nothing wrong with that. I will always support an 'engineering' solution over a hack and these things are worth considering. I'm just not convinced it's an issue that needs to keep you awake at night.
That said, a MOSFET switch, as you originally suggested, would certainly be the more reliable option. With the prevalence of photovoltaic gate drivers, it's pretty easy to rig a pair of NMOS switches with low RDS(on) to do the switching. Just pick a driver/MOSFET combination that ensures adequate VGS to turn the MOSFET fully on.
Tom
Given the use of relays in the automotive industry, the loads switched by these relays, and how reliable the relays are in that application (when was the last time you replaced a relay in a modern car?), I suspect the worry over relay/switch reliability is a bit towards the neurotic end of the spectrum. Nothing wrong with that. I will always support an 'engineering' solution over a hack and these things are worth considering. I'm just not convinced it's an issue that needs to keep you awake at night.
That said, a MOSFET switch, as you originally suggested, would certainly be the more reliable option. With the prevalence of photovoltaic gate drivers, it's pretty easy to rig a pair of NMOS switches with low RDS(on) to do the switching. Just pick a driver/MOSFET combination that ensures adequate VGS to turn the MOSFET fully on.
Tom
OK, we probably have our own ideas over this and on how we would implement any such battery powered arrangement... which is fine
Personally I would always go down the electronic route but my main concern originally was that anyone using batteries should not be connecting battery A (plus rail) and then battery B (minus rail) or vice versa. Which I could see some doing, even if only to test. Make sure both are switched together. Nuff said now
Personally I would always go down the electronic route but my main concern originally was that anyone using batteries should not be connecting battery A (plus rail) and then battery B (minus rail) or vice versa. Which I could see some doing, even if only to test. Make sure both are switched together. Nuff said now
I want to use these SSR for switching and the batteries are loaded with a balancer, so every batterie is on the same level.
https://www.ebay.de/itm/SSR-25DD-einphasiges-Solides-State-Relais-Modul-25A-DC-5-60V
GY/263337760326?_trkparms=aid%3D222007%26algo%3DSIM.MBE%26ao%3D2%26asc%3D43785%26meid%3D16c33c87a51d4b57a74d78deadc43d83%26pid%3D100005%26rk%3D11%26rkt%3D12%26sd%3D262319035219%26itm%3D263337760326&_trksid=p2047675.c100005.m1851
https://www.ebay.de/itm/SSR-25DD-einphasiges-Solides-State-Relais-Modul-25A-DC-5-60V
GY/263337760326?_trkparms=aid%3D222007%26algo%3DSIM.MBE%26ao%3D2%26asc%3D43785%26meid%3D16c33c87a51d4b57a74d78deadc43d83%26pid%3D100005%26rk%3D11%26rkt%3D12%26sd%3D262319035219%26itm%3D263337760326&_trksid=p2047675.c100005.m1851
True. The LM3886 gets rather cranky if you only connect the VEE rail.
Cool! I suggest buying the SSR from an electronic component distributor rather than eBay, though.
Tom
Tom, I never doubted your design. Since this is the first amp I've ever built and you've been working in this field for many years I think it's save to assume that whatever noise problem I'm having is due to my lack of skills and not due to a problem in your design.
Still I'd be interested in solving the problem. Firstly because I'd like the noise to go away and secondly because I'd like to learn from whatever mistake I'm making.
I spent some time yesterday in an attempt to troubleshoot. Nothing helped - but this is what i found out. And no, I did not remove any resistors that are part of the design.
- As mentioned before the only way for current to flow between RCA shield and chassis is this: rca shield to LM3886dr > LM3886DR GND to SMPS86 GND > SMPS86 to chassis. I understand that this is correct and desired.
- i tried different rca jacks and different signal cables, no change.
- i shortened the signal cables and bundled the power supply cables even tighter, no change.
- there is some change no the noise (less) as soon as i touch the speaker cables on the inside of the amp - don't know if that means anything.
- the noise gets MUCH louder as soon as i open the amp chassis. i used my phone to record it. please use the following link to listen to it. as soon as i put the lid back on the chassis the "quality" of the noise stays the same but it's nowhere as loud - it will be reduced by maybe 75%.
NOISE
as always - any help would be much appreciated!
Last edited:
Oh, that's interesting. When you first talked about "noise", I assumed mains hum as that is the most common "noise".
What you are dealing with is RF interference. I bet it'll go away if you turn off your various computers, cell phones, WiFi routers, etc. If you live in a high-density area, you might need your neighbours to turn off their devices as well. Of course, this is not a sustainable situation.
For the debugging, you will need a pair of shorting RCA plugs. So get a pair of RCA plugs, preferably the metal kind, and connect the centre conductor to the shell with as short a wire as you can get in there. Plug these into the amp.
Make sure that all chassis panels are electrically connected to each other. Verify with an ohmmeter and remove any anodization between panels that don't connect.
Make sure the mains ground (yellow/green wire) is connected directly from the mains inlet to the chassis with as short a wire as possible or via the grounded mounting hole (by R2, R3). Verify with an ohmmeter that you have a good connection from the mains ground (measure at the pin of the IEC connector) to all chassis panels.
If the noise persists after you've grounded the entire chassis, it must enter the chassis through the wiring. The input to the amp has an RF filter on it, so I'd be surprised if it enters there, but that is always a possibility. Connect a 100 pF C0G ceramic cap from the RCA centre conductor of the input connector to the metal chassis with as short leads as absolutely possible. Do the same with the RCA shell.
Do the same with the speaker outputs: 100 pF, 50 V, C0G dielectric from each speaker terminal to the chassis with short leads.
That should take care of it.
I'd use solder lugs tied to a screw for the chassis connections to the caps. There's a Keystone lug specified on the BOM for the SMPS-86. Get a few of those with your next Mouser order.
Also: Would you post a build picture that shows the insides of the chassis? I searched back to your original post about the noise issue (Post 312) but did not find a build picture.
Just for the record: My cranky post (Post 355) was in response to the desire by some to "fix" a "problem" by tearing everything apart, leaving the mess for you and I to clean up. You've been very patient and supportive of me in my situation. Thank you. I'll be happy to help you get the amp going. I'm as surprised as you that it didn't work flawlessly the first time, so we'll have to work together here.
Tom
What you are dealing with is RF interference. I bet it'll go away if you turn off your various computers, cell phones, WiFi routers, etc. If you live in a high-density area, you might need your neighbours to turn off their devices as well. Of course, this is not a sustainable situation.
For the debugging, you will need a pair of shorting RCA plugs. So get a pair of RCA plugs, preferably the metal kind, and connect the centre conductor to the shell with as short a wire as you can get in there. Plug these into the amp.
Make sure that all chassis panels are electrically connected to each other. Verify with an ohmmeter and remove any anodization between panels that don't connect.
Make sure the mains ground (yellow/green wire) is connected directly from the mains inlet to the chassis with as short a wire as possible or via the grounded mounting hole (by R2, R3). Verify with an ohmmeter that you have a good connection from the mains ground (measure at the pin of the IEC connector) to all chassis panels.
If the noise persists after you've grounded the entire chassis, it must enter the chassis through the wiring. The input to the amp has an RF filter on it, so I'd be surprised if it enters there, but that is always a possibility. Connect a 100 pF C0G ceramic cap from the RCA centre conductor of the input connector to the metal chassis with as short leads as absolutely possible. Do the same with the RCA shell.
Do the same with the speaker outputs: 100 pF, 50 V, C0G dielectric from each speaker terminal to the chassis with short leads.
That should take care of it.
I'd use solder lugs tied to a screw for the chassis connections to the caps. There's a Keystone lug specified on the BOM for the SMPS-86. Get a few of those with your next Mouser order.
Also: Would you post a build picture that shows the insides of the chassis? I searched back to your original post about the noise issue (Post 312) but did not find a build picture.
Just for the record: My cranky post (Post 355) was in response to the desire by some to "fix" a "problem" by tearing everything apart, leaving the mess for you and I to clean up. You've been very patient and supportive of me in my situation. Thank you. I'll be happy to help you get the amp going. I'm as surprised as you that it didn't work flawlessly the first time, so we'll have to work together here.
Tom
Just a note that if there is an RF pickup problem using a cell phone as a close-proximity audio recording tool is possibly aggravating the problem. Perhaps if you put the phone in airplane mode (cell phone transmission off, WiFi off, Bluetooth off, NFC off) you could consider its presence benign, but I'd be suspicious otherwise.
Forgive me, I'm not familiar with the wiring, and I'm lazy, is there any resistance between the RCA shield and the chassis?
I believe Mixi used the Neutrik connectors I recommend. They isolate the RCA shell from the chassis, which is what you want - at least under normal conditions. The RCA shell is referenced to ground at the PCB. For the best performance, the impedance should be as low as possible.
The common 'trick' of adding resistance between the RCA "ground" and system ground is applied to combat ground loops, thus isn't applicable in Mixi's situtation. It also generally hurts performance rather than it improves it, which is why I recommend a low impedance connection between RCA "ground" and system ground.
One potential cause for Mixi's issues is that one of the wires on the inputs isn't making a good connection. This can happen if the terminal connectors clamp on the insulation of the wires rather than the conductor ... or if one forgets to tighten the screw all the way during assembly.
I'm a bit puzzled by the RFI issues here. I have an amp exactly like Mixi's (2 x LM3886DR + SMPS-86) mounted on a piece of aluminum (~20x20 cm). No chassis. No tricks applied. No issues. I can take a picture when I get home in early July.
Thus, I'd go through Mixi's amp with a fine-toothed comb to check the connections. I'd also check the boards to see if there's a solder joint that's either bad or missing. It's possible that Mixi's RF environment is more contaminated than mine, but that's starting to grasp at straws rather than providing an explanation. Also, I'm located in what's considered 'inner city' of Calgary, so there're plenty of RF emitters around.
Tom
The common 'trick' of adding resistance between the RCA "ground" and system ground is applied to combat ground loops, thus isn't applicable in Mixi's situtation. It also generally hurts performance rather than it improves it, which is why I recommend a low impedance connection between RCA "ground" and system ground.
One potential cause for Mixi's issues is that one of the wires on the inputs isn't making a good connection. This can happen if the terminal connectors clamp on the insulation of the wires rather than the conductor ... or if one forgets to tighten the screw all the way during assembly.
I'm a bit puzzled by the RFI issues here. I have an amp exactly like Mixi's (2 x LM3886DR + SMPS-86) mounted on a piece of aluminum (~20x20 cm). No chassis. No tricks applied. No issues. I can take a picture when I get home in early July.
Thus, I'd go through Mixi's amp with a fine-toothed comb to check the connections. I'd also check the boards to see if there's a solder joint that's either bad or missing. It's possible that Mixi's RF environment is more contaminated than mine, but that's starting to grasp at straws rather than providing an explanation. Also, I'm located in what's considered 'inner city' of Calgary, so there're plenty of RF emitters around.
Tom
Last edited:
Yes, I was just thinking it sounds like a poor connection, it seems to be a very noisy environment. The case is doing its job to a degree. After checking all the connections I would suggest the capacitors between RCA shell and chassis, it's possible the residual noise is entering the chassis and spraying around before getting to the SMPS gnd and chassis
Thanks for your explanation and your advice on this. I'm very determined to solve this problem
Here's a picture of the current wiring situation. The pot is currently just sitting there without a connection as i'm using a preamp at the moment.
As a first step I'll rework the chassis. I'll take it apart and make sure the anodization is properly removed in all relevant places. Maybe I didn't pay enough attention to this when i put it together. Also I have a "fully vented" (link) cover in place as I thought it'd be needed for heat dissipation - which turned out not to be the case as the amp runs much cooler than I thought. Would it make sense to change to a "normal" lid with fewer openings and less "room for stuff to get it" (for lack of a more scientific expression)?
Also I'll make the connection between ground lug and chassis even shorter, taking it straight to the back panel instead of the side panel. Tom, you said to make sure the ground is connected to chassis via the grounding lug OR the mounting screw. In my case it's connected via the green/yellow wire AND the mounting screw. Is that the way it's supposed to be? Just double checking....
This morning I switched off all computers, tablets, phones, tv, etc. basically i shut everything down except for the amp. no success with that.
Tom, going through it with a fine toothed comb sounds good. I'll try to be as methodical as possible and report back.
Thanks for your help!
Here's a picture of the current wiring situation. The pot is currently just sitting there without a connection as i'm using a preamp at the moment.
As a first step I'll rework the chassis. I'll take it apart and make sure the anodization is properly removed in all relevant places. Maybe I didn't pay enough attention to this when i put it together. Also I have a "fully vented" (link) cover in place as I thought it'd be needed for heat dissipation - which turned out not to be the case as the amp runs much cooler than I thought. Would it make sense to change to a "normal" lid with fewer openings and less "room for stuff to get it" (for lack of a more scientific expression)?
Also I'll make the connection between ground lug and chassis even shorter, taking it straight to the back panel instead of the side panel. Tom, you said to make sure the ground is connected to chassis via the grounding lug OR the mounting screw. In my case it's connected via the green/yellow wire AND the mounting screw. Is that the way it's supposed to be? Just double checking....
This morning I switched off all computers, tablets, phones, tv, etc. basically i shut everything down except for the amp. no success with that.
Tom, going through it with a fine toothed comb sounds good. I'll try to be as methodical as possible and report back.
Thanks for your help!
Thought about that too. Surprisingly moving the phone closer or father away (or switching it off) doesn't make any difference. At least I know it's not the phone's fault.