I did some quick but real tests last year on the switching times over in 'tother thread on all this 
http://www.diyaudio.com/forums/solid-state/191449-output-relays-14.html#post2661555
http://www.diyaudio.com/forums/solid-state/191449-output-relays-14.html#post2661555
The Avago ASSR-V621 PV coupler has a switch off circuit already built in.
Off in 30 microseconds with a 1 nF load.
What more do you want?Best wishes
David
Last edited:
@ Cowboy_film_fan:
Figure 14 of the IPB025N10N3 data sheet tells us that it needs approx 30nC to overcome the plateau which turns the switch from conductive to isolating or vice versa.
For two gates this makes 60nC. With a 14uA charging current the switch will operate more than 4ms in linear mode.
Now have a look to figure 3 of the data sheet. Safe Operating Area.
Assuming an inductive behavior of the load, then the SOA for 4ms is just sufficient to protect an 8 Ohms speaker driven from an amp with +/-40V rails.
Already with a 4 Ohms speaker load the design is not bullet proof anymore...
Real life might be slightly more forgiving, because the speaker load will not act fully inductive.
Ignoring SOA typically leads to the situation that one guy says 'it can work' while another guy has to struggle with customer rejections and 'unexplainable defects'.
Figure 14 of the IPB025N10N3 data sheet tells us that it needs approx 30nC to overcome the plateau which turns the switch from conductive to isolating or vice versa.
For two gates this makes 60nC. With a 14uA charging current the switch will operate more than 4ms in linear mode.
Now have a look to figure 3 of the data sheet. Safe Operating Area.
Assuming an inductive behavior of the load, then the SOA for 4ms is just sufficient to protect an 8 Ohms speaker driven from an amp with +/-40V rails.
Already with a 4 Ohms speaker load the design is not bullet proof anymore...
Real life might be slightly more forgiving, because the speaker load will not act fully inductive.
Ignoring SOA typically leads to the situation that one guy says 'it can work' while another guy has to struggle with customer rejections and 'unexplainable defects'.
The Avago ASSR-V621 PV coupler has a switch off circuit already built in.
Off in 30 microseconds with a 1 nF load.
Best wishes
David
I would like to turn ON reasonably fast as well.
I agree that turning OFF is more critical, because you will have the most unfortunate load to the MosFet, when it has to interrupt a burned amp from the speaker.
But also the situation that the MosFet has to switch ON at high music levels should not be ignored. Fortunately here the inductive component of the load is calming the situation and with identical gate drive speed I would guess that turning ON is causing approximately a factor four less stress to the MosFet compared to turning OFF.
These photovoltaic drivers can be a great simplistic solution, just be aware of the resulting limitations.
'It can work', but does not automatically work for all applications.
Simply follow the reasoning of my previous posting about the SOA and check the situation in your application.
Usually you will not need a relay than can easily handle 85Vrms/50Arms or more as shown in posting #5.
Of course the shown Rolls Royce contradicts with commercial aspects.
But here I am for DIY hobby - not at all for publishing cost optimized solutions.
My question was mainly rhetorical because you discussed a switch off circuit when the ASSR already had one - but your answer is very educational!
The FET capacitance is more than I expected. The first idea that occurs is to drive the PV couplers with a simple capacitor + resisitor to increase the current pulse at turn on and reduce the switch time.
This probably needs more care if the FETs are used as rail protection. The on board capacitors could be a difficult load at turn on.
Best wishes
David
I did some quick but real tests last year on the switching times over in 'tother thread on all this
http://www.diyaudio.com/forums/solid-state/191449-output-relays-14.html#post2661555
The measurements are great value for the discussion here as well, because they are showing one key point.
If you use small MosFets like inside the ASSR you can reach reasonable switching times already with small photovoltaic driving currents and consequently reasonable breaking capabilities even with small MosFets.
If you use the same gate drive but just add giant MosFets, your Rds will drop as desired, but the breaking capability will increase by far underproportional due to slower switching.
I used 2 photo voltaic drivers in parallel per channel on my e-Amp protection board to speed things up. If you are looking for straight muting function + speaker short or output device fail protection, switch the return line of the speaker, then you are down to a few 10's of microseconds. However, as pointed out by Dave, this scheme will not protect against a speaker hot lead short to chassis. I have a solution for that, but it adds a bit of complexity so, I take my chances with the scheme I wrote about on my website. With the MOSFETs, you are still well ahead of any EMR solution
In real life handling accidents happen - who wants to have a
burned amp just because of being careless one time?
Furtheron there might not just be somebody, but something .. i.e. overcurrent protections with automatic restart...
But don't bother to much.
It is really up to everybody to decide, what he really needs.
The more extensive circuit of my early postings offer transition times of a few single microseconds with the 2x3xIRFB4321 and allow to handle 100V/100A in a reliable way - for turning OFF and turning ON as well.
For sure that is not the average need, but DIY fun.
Unfortunately not every data sheet does specify this.
But extensive scientific research
has proven the universal truth:Udcmax = pi x thumb x Uacrated at approx. 50% of the rated current
whereas thumb equals 0.07257...
Means - I never had bad luck when using 250V/AC relays for amps with rails up to 57V. When I was young, I could easily live with the implicated uncertainties.
Kiss is what I have in my living room since 5 years.
The rookie amp directly wired to the speakers, no relay.
But my coming class D proto will have a relay. Of course a solid state relay.
Unfortunately the amp partially really happens to become MIUC (or MINC?)
...found that normal integrated comparators have pretty noisy jitter mechanisms - and decided to build my comparators in discrete. Obviously that's an extremely effective approach to shot the component count through the roof.
Class D needs protection otherwise your speakers will burn. Having had various Tripath amps (yes I know not all class D are the same) I learned that:
- disconnecting RCA cables while powered on = disaster
- vacuum cleaning the floor and accidentally disconnecting a speaker wire = disaster
- soldering/touching chips without taking cautious ESD protection = disaster
It made me steer away from Tripath and class D in general. The protection electronics they need may raise the cost from the total amp to that of a reliable good quality class A/B amp. The only one I have left is a brand new never used Sony TA-F501ES new in box but I will sell that one without even trying it out. Went back to linear amps that can have a knock or two
- disconnecting RCA cables while powered on = disaster
- vacuum cleaning the floor and accidentally disconnecting a speaker wire = disaster
- soldering/touching chips without taking cautious ESD protection = disaster
It made me steer away from Tripath and class D in general. The protection electronics they need may raise the cost from the total amp to that of a reliable good quality class A/B amp. The only one I have left is a brand new never used Sony TA-F501ES new in box but I will sell that one without even trying it out. Went back to linear amps that can have a knock or two
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.