SSR for speaker protection?

Another suggestion. Connect T5 collector to D3-D7 junction. C4 discharging will be quicker when AC is lost. (component numbering, post 65)
Check my last schematic, post 58.

Thank you for the suggestion. Here is layout as per post no. 58.
 

Attachments

  • LKA-SPEAKER PROTECT-SCH-rev1.png
    LKA-SPEAKER PROTECT-SCH-rev1.png
    19.2 KB · Views: 941
  • LKA-SPEAKER PROTECT-rev1.png
    LKA-SPEAKER PROTECT-rev1.png
    93.7 KB · Views: 936
Last edited:
Basically there are four items were protecting

1. Output transistors
2. PCB
3. Speaker
4. Transformer

In protecting the 2SC5200 we see that we need to track, the temperature, current and duration of current. This device is capable of 30A for a duration of 300us, so we have to cut that link within 300us for 30A and even faster for higher currents.

As can be seen above detecting whether or not the voltage at our speaker terminal is AC or DC takes a second place to first dealing with the current and the temperature. The quick and dirty solution for handling protection were VI limiters used in many commercial amplifiers including the venerable DC300. Some added fuses to protect the voice coil.

Arduino can take samples at 1us, but by the time we decide whether or not what were seeing is DC or AC, the current has already done us in. One might argue that whether its AC or DC might not matter much as the current https://www.electronics-tutorials.ws/accircuits/rms-voltage.html
 
Last edited:
A bit of history

A couple of high end amplifiers still use VI limiting, however they feature massively parallel output devices so that the VI never need activate even when the kids invited their buddies over while you were away to have a bass boosted party.

The only fuse on such systems is the AC power line fuse. This is what protects everything from the speakers, transformers and copper heavy PCB. The fuse is specifically rated and would be useless without a soft start circuit.

The soft start on these amplifiers feature temperature shutdown. Many under rated speakers have caught fire on these systems not by DC but by AC ;)
 
Interesting thread, have been playing around with using arduino analog pins to check for DC and resolve within 100ms one might say that that is too slow considering above presumably detects and resolves in 380us. The AC lp filter 100k + 10uf will feature a 3.3 volt zener across the capacitor to protect the arduino input, other pins will check temp, over current and short circuit featuring lean code and timer interrupts. output pins will shut off power and trigger speaker protection, plus printing out on LCD the protection event.

Yes the analog domain is faster, however the digital domain is still fast enough.

We use an analog circuit to detect DC which in turn disconnects the SS speaker relay while at the same time pulling an interrupt pin low on an Arduino. The Arduino in turn shuts the amp down and signals trouble. This gives us redundant shut down for speaker relays and amp protection at the same time and much faster reaction time. How to build a 21st century protection board
 
Here's my take on a speaker protect circuit that I designed a year ago or so, basically a discrete pair of SSRs. It has rail failure detection and DC detection for two channels, using mainly SMT. Optocouplers are miniature ACPL-217-500E's, MOSFETs are DMT10H010LK3, which are 100V and 6.5 milliohm. All fits on a board 2" square. Takes the main rails for power and has a test push button.


The detection circuitry is adapted from Doug Self's designs, but instead of driving a relay I drive an LED and two optocouplers.
The L and R grounds are allowed to float a bit w.r.t. each other and the supply ground.



Note the schematic needs some improvement, I've got the wrong symbol for the zeners (they are labelled with voltages though), and the wrong part number on the MOSFETs.



The key insight is how I float a zener/capacitor 12V supply between the rails with high value
resistors, which need to be sized to provide enough current for the resistor across gate & source when the opto-coupler is on.


The cross-connected 470k resistors are to give hysteresis to the switching, but I haven't yet tested or populated those, the basic protection function works without them.

speaker_protect.png
 
I have been working on another version with smd and tht both and here are the views.
The smd on bottom and also small THT are populated from bottom side.
All electros , transistors from top.
The mosfet is a dual channel smd mosfet SQJB00EP.

regards
Prasi
 

Attachments

  • composite.png
    composite.png
    71.7 KB · Views: 1,028
  • bottom side.png
    bottom side.png
    51.7 KB · Views: 987
  • top side.png
    top side.png
    52.4 KB · Views: 396
  • sch.png
    sch.png
    13.1 KB · Views: 488
Last edited:
how fast are you able to detect ac or dc

There's a couple issues you will have with a design like this. First, it's way too slow. The cone of the speaker would already be against the stop by the time that thing saw DC. Your detection needs to be happening in the microseconds realm, not milliseconds. Second, you have digital and analog circuits connected together, it's going to be noisy. If you are going to monitor two channels you'll have ground loop issues as well.

I've found the best approach is to use a good old fashion analog DC detection circuit connected to the input of a high speed optoisolator. Use the output from the optoisolator to change the state of an interrupt pin on your microcontroller. I use a dual input optoisolator and have one output connected to the microcontroller and the second directly disconnecting the SS relay allowing for a redundant shutdown of the SS relay with the microcontroller latching it off, meanwhile the microcontroller shuts down the amp at the same time minimizing further damage to the failed output stage.
 
I have been working on another version with smd and tht both and here are the views.
The smd on bottom and also small THT are populated from bottom side.
All electros , transistors from top.
The mosfet is a dual channel smd mosfet SQJB00EP.

regards
Prasi

gerbers as per above design. Not tested.

regards
Prasi
 

Attachments

  • SSR_LKA_REV3-smd_2019-04-30.zip
    342.2 KB · Views: 254