Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Pass Labs This forum is dedicated to Pass Labs discussion.

A possible approach to adding a silent start/shutdown to the ACA
A possible approach to adding a silent start/shutdown to the ACA
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 14th September 2018, 09:40 AM   #1
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
Default A possible approach to adding a silent start/shutdown to the ACA

One quirk of the hugely popular ACA that seems to regularly get a mention relates to the strange noise it can make at power on and the loud thump that occurs at switch off. This kind of behaviour is actually 'normal' for pretty much any AC coupled amplifier. AC coupling here relates in the main to the large speaker coupling capacitor and the fact this has to both charge and discharge via the load (your speaker) when power is applied and removed.

It is also possible (you would have to confirm or otherwise) that the SMPS may 'ramp' up the voltage due to momentary current limiting becoming active at power on.

At this point I have to say I have never built an ACA and so ALL that follows is 100% theoretical, however those interested enough to delve a little deeper may find some ideas in amongst all of this on how to make it virtually silent.

A graphical representation shows the problem. Here we have a simulation of an ACA showing the sequence of events as power is applied. The supply voltage and amplifier output voltage (the black speaker terminal) are shown with reference to the left hand voltage scale. The load current (your speaker) is referenced to the scale at the right. We can immediately see the large switch on thump and an even larger switch off related event. The ACA has no signal applied during this test.

A possible approach to adding a silent start/shutdown to the ACA-aca-1r-jpg

Is it possible to eliminate this and give a silent start up and shutdown ?

Using a speaker relay would be considered one possible route, however relays are a little power hungry and can suffer from contact degradation over time. This isn't some imaginery issue that happens to other people... its happened to me... and as witnessed by many threads on these forums it also happens to many others as well.

One real alternative is a solid state relay contstructed from two back to back power MOSFET's. These are becoming more and more popular because they suffer no degradation in use, and have a current switching capacity limited only by the FET rating. Voltage rating of this type of relay is again determined by the FET. The series resistance (what would be contact resistance in a traditional relay) is twice the value of the Rds of the chosen FET.

Downsides of the FET relay.

In practice when used for speaker switching duty, one real problem is that of providing the necessary gate drive voltage. The reason for this is that the drive voltage would normally have to be 'floating' and not referenced to ground or the supply rail.

The following explores the FET relay in more detail and the diagram below is a practical relay constructed from two N Channel power FET's. The key to the relays success relies on the drive voltage V1 being floating. If that requirement is met then the relay behaves as expected and is truly bi-directional, just like a mechanical one. In this form the relay is suitable for speaker switching duty in both AC coupled amps and also DC coupled running off a dual supply.

The floating drive voltage is traditionally derived from a 'Photovoltaic Coupler' which can be thought of as being like a tiny solar cell that is illuminated by an LED. When the LED lights, the cell generates voltage (typically 7 volts), and this voltage is applied as a floating source between gate and source to turn the FET's on or off. The whole thing is often drawn in a similar way to a standard opto coupler and packaged in the same manner. The on/off isolation of this type of relay is excellent at DC and low frequencies, and becoming worse at high frequency due to the junction capacitance of the FET's. In practice this is a non problem for our application because the load being switched is (relatively speaking) of very low impedance.

A possible approach to adding a silent start/shutdown to the ACA-aca-2r-jpg

Turning now to the ACA, we have the option of placing such a relay in either the speaker positive or speaker negative feed. It is simply a series switch and so in that sense it doesn't really matter where in the circuit it is placed.

Before considering what may be possible we first have to look in detail at the problem we are trying to solve. Although it may seem obvious when mentioned, one real factor to be considered is that when the speaker is not connected, the time constant of the speaker coupling cap and R14 (whcih is the caps only means of charging) is large.

We can see that below. Here the speaker is not connected.

A possible approach to adding a silent start/shutdown to the ACA-aca-3r-jpg

There is no easy way around this long time constant. Technically complex solutions would be to employ a low value shunt resistor to give a super quick charge time, the resistor then being switched out of circuit. In the ACA about the best we can do is to reduce the value of R14 and that is what is done here. Reducing by a factor of ten gives us something more realistic to work with, and now we could reasonably say that a delay of 20 to 25 seconds should give a much reduced switch on thump. The longer the delay, the quieter its going to be. Given the low power output of the ACA a 1 watt 100 ohm is more than adequate.

A possible approach to adding a silent start/shutdown to the ACA-aca-4r-jpg

So having come up with a suitable solid state relay, and now having some idea of the delay needed the next step in the design process would be to construct some form of drive circuit for the FET relay. The ACA does us some favours here because of the fact it runs on a tightly regulated supply with very defined start up and shut down characteristics.

(I know the start up and shut down behaviour of the SMPS will be tightly defined, the trouble is that I don't know exactly what the numbers might be with regard to supply rise and fall times and so I'm going to have to make some assumptions)

A simple comparator seems a likely possibility to begin the design and only one is needed because both ACA channels would share the drive circuitry.

So firstly we have to set up our supply rail in the simulation and for this I am choosing a rise time of 200ms and fall time of 150ms. Next we add the comparator and set an R/C time delay on the inverting input. The non inverting input is fixed at 20 volts by the zener diode.

The operation is as follows. When the supply voltage appears, the RC network delays the rise in voltage at the comparator input. It is the classic exponential charge curve. When this rising voltage crosses the fixed level provided by the zener we see that the comparator output flips from low to high. This voltage can then be used to operate our relay. The function of the diode is simply to dump the charge in the timing cap back into the supply rail when power is removed. This allows the circuit to be instantly ready to work again if needed.

A possible approach to adding a silent start/shutdown to the ACA-aca-5r-jpg

A possible approach to adding a silent start/shutdown to the ACA-aca-6r-jpg

Putting it all together.

The following shows the relay in the off state (that floating voltage source between G and S is zero). The upp er right voltage source represents the ACA and is a 4 volt peak sinewave (1 watt/8ohm) riding on a DC voltage of 12 volts. The plot shows the load current (leakage) which is pretty small.

A possible approach to adding a silent start/shutdown to the ACA-aca-7r-jpg

Now we change the floating voltage source to 12 volts which will fully turn on the FET's. Again we look at the load current and can see that the FET's behave as if they were essentially a wire link or short. A near perfect relay.

A possible approach to adding a silent start/shutdown to the ACA-aca-8r-jpg

This shows the voltage at the 'earthy' end of the load, in other words the voltage lost across the relay. The lower the Rds value, the lower the losses.

A possible approach to adding a silent start/shutdown to the ACA-aca-9r-jpg

(Audiophile hint using FET's with a higher Rds has the effect of reducing the damping factor. Heretical audiophile hint Using dissimilar FET's might be felt to introduce a very low level 'pleasant' even order distortion profile to the resultant current passing through the relay)
Attached Images
File Type: jpg ACA 1R.jpg (70.7 KB, 705 views)
File Type: jpg ACA 2R.jpg (50.2 KB, 688 views)
File Type: jpg ACA 3R.jpg (68.7 KB, 684 views)
File Type: jpg ACA 4R.jpg (67.4 KB, 686 views)
File Type: jpg ACA 5R.jpg (91.7 KB, 696 views)
File Type: jpg ACA 6R.jpg (83.3 KB, 693 views)
File Type: jpg ACA 7R.jpg (110.0 KB, 686 views)
File Type: jpg ACA 8R.jpg (112.0 KB, 690 views)
File Type: jpg ACA 9R.jpg (89.6 KB, 684 views)
  Reply With Quote
Old 14th September 2018, 09:40 AM   #2
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
This is a distortion plot of the current passing through the load. The first example is with similar FET's, the second with dissimilar.

A possible approach to adding a silent start/shutdown to the ACA-aca-10r-jpg

A possible approach to adding a silent start/shutdown to the ACA-aca-11r-jpg

Getting it all to work.

This is where it gets more difficult. The photovoltaic coupler is the sure-fire way to get a truly floating drive voltage, the downside being the cost... although of course in the scheme of things its all relative. They are not that expensive. We also have the problem of how to simulate this. I came up with the idea of using a traditional optocoupler and a fixed 12 volt source. Lets see how that works out.

This shows the idea in action. The audio is muted until the comparator changes state, at which point the relay closes and audio is passed. The third image shows a close up of the moment the relay turns on. A practical relay using a photovoltaic coupler would need no other components apart from the coupler and FET's. The 1meg in the simulation is to stop the tiny leakage of the opto coupler model from interfering with the off state of the relay.

A possible approach to adding a silent start/shutdown to the ACA-aca-12r-jpg

A possible approach to adding a silent start/shutdown to the ACA-aca-13r-jpg

A possible approach to adding a silent start/shutdown to the ACA-aca-14r-jpg
Testing the idea with the actual ACA simulation.

This is where issues come to light, however, these may be more down to the simulation than a real problem. Specifically, the main problem is seeing a switch off thump (current in the speaker) when the relay should be off. I know from having used these relays that they behave as expected and so I'm left wondering if the issue here really is simulation related.

At this point we take a step back and redraw the circuit with a view to clarifying and simplifying. A different model for the opto seems to obviate the need for the 1meg resistor. In a real build it wouldn't matter anyway because the photovoltaic coupler is all that is needed. The comparator (LT1017) has an active class B output stage, however the current source ability is limited. This means we still need the traditional 'pull up' resistor to supply the LED in the coupler. A zener has also been included in the feed to the LED and this is purely to make absolutely sure the relay can not be forced into conduction due to any small glitches/transisions etc that may momentarily occur on the comparator output.

This shows our basic working circuit, and following on from that the voltage and currents at critical points in the design. You can see the supply voltage at the point of switch off together with the comparator output voltage and current in the Zener/LED feed. Note the speaker current. The small current pulse at around 12 seconds is solely due to the time delay chosen before the relay activates and 12 seconds is not long enough for the both the amplifier and its 3300uF speaker coupling cap to charge via the 100 ohm resistor (R14). How long a user would want to wait before the speaker was activated is down to personal choice.

A possible approach to adding a silent start/shutdown to the ACA-aca-15r-jpg

Remember it is not just the 3300uF cap that is the problem here. The amplifier itself has a very long settling time, realistically around 18 seconds from power being applied as can be seen here.

A possible approach to adding a silent start/shutdown to the ACA-aca-16r-jpg

Now to the 'big' problem that may well just be a problem with the simulation. This shows the moment of switch off. Note the way the 24 volt rail collapses (one of my assumptions outlined at the start... I hope I have picked a realistic rise and fall time for the 24 volt rail), and in particular note how quickly and cleanly the comparator flips state removing drive current from the photovoltaic coupler. Once the drive current is removed, the relay is off and this is the puzzle/problem... the simulation goes on to show a current spike in the speaker that occurs after the relay is off. I have no answer to that besides saying that it shouldn't be doing it.

A possible approach to adding a silent start/shutdown to the ACA-aca-17r-jpg

The big question. Is this current spike real or not ? If the relay were mechanical and driven by the comparator output voltage then the answer is a definite no, the relay must be off at that point in time and so no current can flow.

A real and practical solid state relay behaves like a mechanical one. So where does this spike come from ?

To elaborate on this further I took a plot of the current flow in the FET source lead and overlaid this with the relay on/off control voltage. The voltage on the LED assumes a slightly negative value... puzzlement over that... and so at this point I'm going to say that the behaviour of this in simulation is possible erroneous and that the a real photovoltaic coupler would work as expected.

A possible approach to adding a silent start/shutdown to the ACA-aca-18r-jpg

So job done then, at least as far as a theoretical exercise is concerned. Well not quite... the photovoltaic coupler is a main cost item and so I wondered if it could be eliminated. The fact the ACA is single rail, and the fact the relay is in the negative speaker return suggests it might be possible. So lets see.
Attached Images
File Type: jpg ACA 10R.jpg (94.4 KB, 692 views)
File Type: jpg ACA 11R.jpg (100.2 KB, 687 views)
File Type: jpg ACA 12R.jpg (103.3 KB, 690 views)
File Type: jpg ACA 13R.jpg (68.0 KB, 687 views)
File Type: jpg ACA 14R.jpg (106.0 KB, 691 views)
File Type: jpg ACA 15R.jpg (113.1 KB, 692 views)
File Type: jpg ACA 16R.jpg (73.5 KB, 694 views)
File Type: jpg ACA 17R.jpg (74.0 KB, 691 views)
File Type: jpg ACA 18R.jpg (77.5 KB, 688 views)
  Reply With Quote
Old 14th September 2018, 09:41 AM   #3
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
I have to say I was amazed at the result. Can we trust this ? The switch off pulse is minimal and the circuit very simple.

A possible approach to adding a silent start/shutdown to the ACA-aca-19r-jpg

A possible approach to adding a silent start/shutdown to the ACA-aca-20r-jpg

So finally here are the two versions to compare, each tested at no signal and full output. That large switch off pulse looks bad and yet everything says it shouldn't even be there with the Photovoltaic version, this method of using a floating voltage source (the photo voltaic coupler) being the 'gold standard'.

A possible approach to adding a silent start/shutdown to the ACA-aca-21r-jpg

A possible approach to adding a silent start/shutdown to the ACA-aca-22r-jpg

And the second version. Is it to good to be true ? This is the version that I would expect to be fussy on implementation... however the simulation says otherwise.

A possible approach to adding a silent start/shutdown to the ACA-aca-23r-jpg

A possible approach to adding a silent start/shutdown to the ACA-aca-24r-jpg

So that just about wraps up the theoretical side of this, the next step for any of those interested would be to actually build one and see. The design is easy to implement to try and test and can be literally 'tagged onto' the speaker output terminal with only a 24 volt feed needed.
Attached Images
File Type: jpg ACA 19R.jpg (110.7 KB, 693 views)
File Type: jpg ACA 20R.jpg (78.8 KB, 688 views)
File Type: jpg ACA 21R.jpg (67.2 KB, 685 views)
File Type: jpg ACA 22R.jpg (77.5 KB, 685 views)
File Type: jpg ACA 23R.jpg (65.4 KB, 677 views)
File Type: jpg ACA 24R.jpg (73.1 KB, 679 views)
  Reply With Quote
Old 14th September 2018, 09:41 AM   #4
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
These are the .asc files for anyone wishing to see how the simulation behaves. These files were created using default models included in LT XVII
  Reply With Quote
Old 14th September 2018, 09:54 AM   #5
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
Hope those images load OK as the forum software hasn't scaled them all equally.

Edit... I'll try and sort the images later today hopefully.

Edit 2... All images replaced with 1024*xxx pixel size. They now seem to load correctly and not overspill the margins.
  Reply With Quote
Old 14th September 2018, 07:11 PM   #6
takitaj is offline takitaj  United States
diyAudio Member
 
Join Date: Apr 2012
Location: Maine
Hi Mooly. I'm still deep in the learning stage so forgive me if this is a dumb question.

I was wondering if the delay in the mosfets shutting off could be due to the capacitance charge of the mosfets themselves? Could that be why it improved when you added the resistor between gate and source? Would a resistor to ground somewhere help at all?
  Reply With Quote
Old 14th September 2018, 07:37 PM   #7
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
Its a good question and your reasoning could be correct

I tried quite a few things with the simulated photovoltaic coupler but I will re-run those simulations later and have a play. The real photovoltaic coupler (which I can't fully simulate) actually has an active turn off circuit within it.

The simple version with no coupler seemed to work better than expected and that included a 100k from gate to source to both eliminate any issue with stray pickup from what is a high impedance node (as its connected via a Zener) and it also aids removing charge.

It is an interesting effect though.

This is how the real coupler looks in block diagram form.
Attached Images
File Type: jpg Capture.JPG (40.0 KB, 63 views)
  Reply With Quote
Old 14th September 2018, 08:39 PM   #8
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
I rigged this up to try and see how the FET's behave when switched off. The switch supplying the gate has a 'nearly' infinite off resistance at 1000 Meg ohm. The 10 meg from gate to source discharges the gate capacitance.

Making the resistor 100k sharpens things up no end.
Attached Images
File Type: jpg Capture.JPG (173.7 KB, 73 views)
File Type: jpg Capture2.JPG (212.9 KB, 33 views)
  Reply With Quote
Old 14th September 2018, 08:39 PM   #9
Preamp is offline Preamp  Germany
diyAudio Member
 
Preamp's Avatar
 
Join Date: Jul 2008
Location: Monkey Island
Just a quick thought while skimming through your thorough work: What about putting that solid state relay in parallel with the speaker and shunt that thump to ground? I for one would not like to wait half a minute for my amps to come on... (well, I'm already annoyed by two seconds turn-on delay )
  Reply With Quote
Old 14th September 2018, 08:51 PM   #10
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
A possible approach to adding a silent start/shutdown to the ACA
Well... a shorting relay is certainly another possibility.

I'm not sure how it would work in practice with a solid state relay because the circuit needs voltage to operate the relay, and voltage means the speaker cap is already charging. That would be a whole new design exercise I think.

A mechanical shorting relay would of course be closed at the crucial time.

It's an interesting idea though.
  Reply With Quote

Reply


A possible approach to adding a silent start/shutdown to the ACAHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
P3A amp shutdown 'plop' jacques antoine Solid State 11 28th May 2015 08:23 AM
NTC thermal shutdown lazzer408 Power Supplies 36 14th March 2015 06:59 PM
Squeal on shutdown famousmockingbird Tubes / Valves 48 11th February 2014 07:55 AM
OPA548 and shutdown kafka Chip Amps 9 5th September 2004 05:26 PM
Shutdown noise energyandair Solid State 3 10th May 2003 11:52 AM


New To Site? Need Help?

All times are GMT. The time now is 08:59 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki