DC speaker protection for Sure class-D amps?..
I was going to post this in the "4*100 Watt Sure" thread, but I didn't want to hijack it with my ramblings. OK, here goes :) ...
My 4*100 Watt board has arrived and is sounding great. It only took 6 days to arrive from Sure (China) to the UK!
I'm currently testing it with some 1uF M-caps on the inputs and I'm very impressed with the sound. The stock caps (on the phono inputs) sounded quite "wolly" in comparison.
With the M-caps, the sound is crystal clear, but the frequency response still sounds very well balanced. The amp has a lovely amount of "kick" to it, and the mid-range sounds great on vocals. Personally, I think this amp already sounds a tad better than my RevC chipamps, and I'm currently only running the Sure amp using a crappy 20V 3A laptop PSU.
It's all looking (and sounding) good so far. The thing is, I'm planning to use two Sure amp modules in my DIY build (7.1 with DSP etc.), but I'm worried about hooking up my main speakers (M&K S-150's) to the Sure amp without any DC protection relays (especially since a couple of people on here have seen the "magic smoke" escape from their amps already).
I've searched the forum, but can't seem to find a DC protection circuit which is specific to the way the Sure amp is configured. If anything bad happens to my DIY amp, I'd MUCH rather the amp gets fried than my speakers, so I want the circuit to be very reliable...
I've knocked together a simple DC protect circuit on veroboard for testing, but I'm confused about the best / simplest way to implement this on a class-D type amp? My circuit seems to work fine for driving a relay directly and it disconnects the speaker when DC (or very low bass) is present. The problem is, I don't know how foolproof the circuit is?
I've also attached a diagram of one of the DC protect circuits I found floating around the web.
My circuit is based on the first (DC protect) sections of the attached diagram, but with a resistor added to the second speaker input wire and with only one capacitor on the OUTPUT of the diode bridge. (I'll have to post a diagram at some point to make this clearer.)
What I realized with the Sure amp is that it still needs the bridge rectifier to be put across BOTH speaker wires because the amp is basically configured like a full bridge (I think?).
So, neither of the speaker outputs will have a negative voltage on them (with respect to ground), but they can still "swap polarity" in response to the input audio to achieve the output voltage swing (is this a correct way of describing it?).
Are there any ways of simplifying the DC protect circuit due to the fact that the neither speaker wire will go negative, or would something like a diode bridge always be necessary? I've searched the forum and have seen people talk about using a differential pair of transistors for DC detect, but how would this be implemented on the Sure amp? I know a few commercial amps which use a transistor pair, but I don't know how this could be applied to the Sure amp.
In the attached circuit, am I correct in saying that the emitters of Q5 / Q6 will need to be grounded, else they won't be able to pull the output signals low? Is there a good reason to use two capacitors on the protection inputs (to make a non-polarized cap) rather than a single cap placed AFTER the diode bridge? I know the diode bridge will produce a small voltage drop, but will using a single cap be a disadvantage to the circuit's operation?
Also, although this circuit has a diode bridge, it looks like it's still designed for amps which have one wire of the speaker output at 0V. When using the Sure amp with this circuit I think it would be a very good idea to add another resistor on the other speaker input wire, otherwise one "leg" of the amp will be connected to ground directly through the diode bridge!
(as a side note, some of you might have noticed that the LEFT channel input has a resistor but the RIGHT channel only has a resistor on the 0V wire?)
I'm only really interested in the DC protection parts of the circuit as I don't think the Sure amps require much of a turn-on delay. Although, it would be fairly simple to add a separate supply for the relays to eliminate turn-off thump from opamps and DACs if the AC power is removed.
Anyone have any thoughts on this type of protection circuit?
Actually, I just realized that putting the cap AFTER the diode bridge would mean that it is no longer an RC circuit.
On my circuit with only one cap, it only responds to fairly high voltages. The circuit didn't even trigger with a 4V battery across the input, so that can't be good.
I'll change it back to using two caps (non-polar) after the input resistors and see how it responds and I'll report back.
Speaker protection module
I would just use one of the many ready made speaker protection modules that are available.
Speaker protection Board - eBay (item 190348617421 end time Nov-13-09 20:14:56 PST)
Nice work on the inductor replacements btw, I enjoy reading up on your mods. :cool:
I've looked into those boards before, and I know the UPC chip is a proven design, but do you think it would work properly for the Sure amp?
The root of my confusion is that the diagram in the UPC datasheet appears to show that the chip is designed mainly for amplifiers which have one of their speaker wires fixed at 0V (and the other wire usually goes both positive and negative). So, it can only detect DC on ONE speaker wire (referenced to ground) and not across the speaker wires?
This is true of many of the other DC protection circuits on the Web. Although, when a diode bridge is used, it should in theory work OK on the Sure amp?
(The protection modules on eBay suggest that they are "great for T-Amp's", but does this necessarily mean they will work on a bridged amp?)
Please let me know if the following is correct because I'm still trying to get my head around it all.... :confused:
The Sure amp is configured as a full bridge, so if you wired it to a UPC chip like in the attached diagram, it can detect voltage on only one of the speaker wires at a time (it only shows one input resistor per channel).
I just tested the outputs of the Sure amp to confirm that each speaker wire output is actually at half the power rail voltage with no input signal. So, for my 19V power rail, both the positive AND negative speaker outputs are at 9.5V (when referenced to ground) with NO music playing...
This obviously gives a net voltage of 0V across the speaker (apart from the small DC offset of an unmodified amp :) ).
I don't think it would work to simply connect a resistor from both speaker wire outputs to the UPC chip because when both speaker wires are at the same voltage (say 9.5V again), the speaker will see 0V, but the voltage out of this resistor network will still be 9.5V when referenced to ground, and this will trigger the protection IC.
Attached is the diagram from the UPC datasheet - it shows how it would connect to a normal push-pull amp, but this type of amp only needs one of the speaker outputs to be monitored (the other is at 0V) - how would this work for a Sure amp when both "plus" and "minus" outputs are active??
I guess the solution would be to add another resistor (to the "minus" output wire), then add a diode bridge between the other ends of the resistors. The DC minus output of the bridge will be grounded, then the DC plus output will go to the UPC chip input. I think this should work OK, but I'm not 100% on this sort of stuff.
I just dropped the voltage using two resistors into the PIC i/o one pair for +ve phase and anotehr set for the -ve phase.
If the output stays >20 volts or < -20volts for 500mS then I drop the relay.
Using the PIC also means I can have a 3 second anti thump power up delay.
I use this circuit on my two disco amps and never blown a speaker due to DC.
I accidentally shorted the output once and blew the output transistor4s but the DC protect jumped in to save the speakers.
I had thought about using a PIC once I have all the basics up and running.
I've actually used PICs a lot with various projects, but the irony is that I do like the idea of using "old skool" electronics for the DC detect part.
I want something that will be super simple yet very reliable. Then again, using a small PIC with an internal clock should be fairly simple and would work OK as long as the supply is stable.
You would still need a way of determining the voltage across the speaker wires on the Sure amp though. It seems that a diode bridge is still the simplest method....
I've attached a diagram of the circuit I'm testing atm. Whether you put a 16uF cap before the bridge (non-polar), or a 33uF cap after the bridge (polar) the circuit appears to work in a similar way. The placement of the cap only appears to change the frequency response slightly - the relay will switch off with a sine wave of below 1.2Hz with 16uF and below 1.4Hz with 33uF.
Another issue is that, at low frequencies, the circuit needs around 8.5 volts across the speaker terminals before the relay will drop out! This bothers me because if the amp fails in a way which presents only 8V of DC across a 4 ohm speaker, this is still 2 AMPS of current! :hot:
OK, this next bit of rambling is quite off-topic, but I thought it might be of interest. Please let me know if I'm rambling too much again, I don't post very often! ;)
I'm actually using the CPU and DSP boards out of a Denon receiver to decode 7.1 DTS / DD etc. So, if I can work out a reliable DC detect circuit for the Sure amp, I could use the orignal /PROTECT signal on the Denon CPU board to shut down the relays.
Atm though, I'm not comfortable with the way the DC detect circuit is working and I think using a PIC chip is quite a bit of hassle for something I'd imagine has been built a few times before (DC detect for bridged class-D)? Again, I did search the forum but I couldn't dig up much.
Basically, using a diode bridge should work in a very similar way on the Sure amp (no negative swing vs GND) as it does on a push-pull amp, but the fact it takes over 8 volts to trigger the protection is quite a scary thought to me.
As a side note: the DSP will eventually connect to my 8-channel Sabre DAC board (courtesy of NeoY2k) once I get the parts in. The DSP was previously running fine with in 2-channel stereo via my Buffalo DAC, but I wanted 8 channels.
I'm planning to build a controller with LCD UI for the amp from scratch. I need to figure out how the original CPU configures the DSP though. Once I've done that, I can control the DSP / DAC / volume / protection / inputs all from my own UI (with remote control).
Anyway, I digress (big time) :o ...
Does anyone have any ideas for reliable DC detect circuits for a bridged class-D? I'll look into the PIC idea though as I can simulate this before building.
Thanks for reading.
Sorry, I'll have to correct the part about the activation voltage... Under the Proteus sim, the signal generator doesn't produce anywhere near the selected voltage peaks. You also have to turn it down to 0Hz just at the right moment to "catch" it on a peak.
The peak voltage is only around 2.5V when the dial is set to 8V, so I don't think it's even related to RMS?
Anyway, I since replaced the signal gen with a pot and the results are more promising. The circuit only needs 2.2V of DC to turn off the relay.
The relay still drops out with around 2.2V whether you use a polar or non-polar cap (at least on the sim), but it will affect the corner frequency / response time slightly.
Could 2V of DC across the speaker cause much damage though? This is still around half an Amp on a 4 ohm speaker? (assuming a failed amp will still supply that much current).
This is why I hate it when people "tap-out" the speaker phase using a 1.5V battery. It must be doing some damage if it's held on for too long?
Attached is a better example of the circuit. The software doesn't show the forward voltage drop of the generic diode bridge, so there are quite a few variables to play with.
Project 111 - PIC Based Speaker Protection
This one could probably be used on the Sure amp if the inputs were modified with diode bridges.
|All times are GMT. The time now is 01:43 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Copyright ©1999-2016 diyAudio