Transcendent Sound Forum
Latest Topics
 
 
 


Reply
  Author   Comment   Page 2 of 6      Prev   1   2   3   4   5   Next   »
Wolfgang
Reply with quote  #16 

I have some questions about solid state relays (SSR). I have never used them with audio signals before.

Is there anything to be considered regarding transferring the full audio bandwidth? I couldn’t find any information in the data sheets?

Using an AC SSR ...would it interfere with the function of the SSR when/if the amp shows DC at the output (little DC offset with PP amps or with SE amps in the case of clipping)?

What ratings for the SSR would be a good choice for low power OTLs (1-15W), like 2A/10V or even higher in order to be always far below the max ratings? A-T not more than 1:2.

What about the short voltage peaks at the output when the amp is turned on (1-2 sec) and stabilizes? Do they have to be considered in the context of the max ratings of the SSR?

Thanks for every input!

Kelvin Tyler
Reply with quote  #17 
Wolfgang,
I take it you are considering back-to-back MOSFET AC type switches in this application.
Let me say straight away that my experience with these devices is very limited, so please do not take the following as authoritative in any way. Just thinking aloud as it were ( sticking my neck out!).
I have used the opto-coupled LCA715 device, but not in an audio application. This device is rated on the output side at 2A/60V and switches on and off in less than 1msec, so it might be OK with a low power amp. I think the bandwidth should be adequate and I do not see any problems with amp switch-on transients if you are using a 300B DC SE amp. It is simple to limit output voltage excursions with an MOV or back-to back zeners.
Just for you amusement, I am including an 'in principle' sketch of what I imagine you have in mind. Hardly a fleshed out design, and I still have reservations about switching reactive loads in this way. The control logic would be simple enough to implement, and you could easily switch the speaker out of circuit if an amp fault happened.
The LCA715 devices are very inexpensive so it would be simple to try one for suitability without implementing full switching logic. Unfortunately, I do not have any to hand at this time myself.

switch.jpg 
Kelvin

Wolfgang
Reply with quote  #18 

Thanks for the input, Kelvin. It is always helpful to get your opinion. I have ordered some SSRs that could work including  the  LCA715. First step is testing SSRs in the signal path.

Switching “reactive loads” will be interesting and challenging. I hope somehow that very precise and fast switching will make it work. My biggest concern is some possible distortion from the MOSFET at zero crossing.

I don’t see any alternative to  SSRs (very high Ri-o) because the core of the circuit which I am planning to use for setting input sensitivity, threshold, attack and release time is digital. Digital is on the other hand the best way to set the necessary parameters precisely enough. Please take a look if you are interested:

http://www.mectronicsinc.com/projects/vm-ai1i/

Unfortunately I need 3 switches to make this work like in your drawing. I wished I could avoid SW1 and could have a pure-OTL connection without any switching.

Achim
Reply with quote  #19 
Wolfgang,

when you switch in the autoformers, they have some kind of "startup sequence", probably the period where the core gets magnetised. If your circuit is not able to have a base magnetisation on the autoformers at all times, it will not be worth the effort IMHO.

Who knows, best of luck!

PS: As to your question about mains quality: I am currently looking for the first time at my power supplies with a scope - the voltage coming from the mains is terribly crippled. Also I am amazed and terrified by how much EM emission there is from the power cable even when the device is off! I will look for shielded power cabling. Inside Bruce's amps he uses shielded cable from the inlet on the back to the switch. But having these emissions in the 100mV range right in front of the back door is unsettling.  
Kelvin Tyler
Reply with quote  #20 
Achim. Yes, I tend to agree. It might be possible to omit SW2 on my putative sketch, but I am not sure what would be the effect of the A-T inductor's presence on things.

Wolfgang. You have chosen what I would call a 'Rolls-Royce' solution with the Mectronics VU device, although I see it is not that expensive. I had in mind something a bit simpler. A similar analogue front end, amplifying and rectifying the signal, and then triggering an adjustable voltage comparator with a bit of low-pass filtering. This could be done with a couple of op-amps. The 2 state, 2 or 3 output digital control logic could then be implemented with a single TTL hex inverter (7404 or 7414) or equivalent CMOS device. Maybe you would prefer greater overall control.  It is important that the switch action is rapid as the MOSFETs dissipate most when switching.
Looking at the LCA715 spec sheet, I do not see any cross-over problems. Time will tell.
Cheers,
Kelvin
Kelvin Tyler
Reply with quote  #21 
CORRECTION. Sorry, should have written SW3, not SW2, in last post!
Kelvin
Wolfgang
Reply with quote  #22 

Achim,

no doubt it looks pretty bad on a scope. But people argue it will be  rectified and filtered anyway so why bother..but there is more to the story. Don’t get me going…

Valid point with the “start-up sequence” of the A-T but what to do? How did you find out about it with your A-T either always connected or disconnected?

To keep the A-T always in the circuit must have some audible effect compared to pure OTL. Do you think it would be possible to feed some sine signal from outside the audible frequency range to one of the unused connections of the A-T that would keep it always “on” ? Would that solve the problem?

 

Kelvin,

We will need the CMOS switch anyway because of SW1. Otherwise I think the SSR #2&3 could be switched directly (some minor adjustment) with the voltage from the LED meter. I would be extremely interested to see your original circuit and how you wanted to make the adjustments precise enough. I couldn’t understand from your description how you would be able to choose between different thresholds and adjust parameters in general. But I don't need a Rolls Royce if there is a simpler way.


Achim
Reply with quote  #23 
Wolfgang,

I noticed the startup sequence with the ATs connected between my Pinnacles and the speakers. Mine have 2-3-4-6-8 ohm taps on speaker terminals, so you are free to transform impedance up or down as seen from the amp. Amp connected to 8 ohms and speakers on one of the other taps transforms up, amp on one of the others and speakers at 8 transforms down.

If I remember correctly I noticed the startup sequence as some sort of whine through the amps at amp startup. I also noticed movement of AT stored energy as sound in the speakers when changing taps with amp on.

I had the ATs connected at all times - as opposed to the galvanically isolated tapping which my ATs also offer, there was *almost* no degradation of the sound. That slight bit only became noticable with very phase accurate speakers.

Cheers,
Achim

PS: Yes, AC is a topic for elsewhere.
Wolfgang
Reply with quote  #24 
Achim,

I obviously misunderstood your earlier post. Now it rather looks like the opposite: it's not that the core would need to be "magnetized" which would make some kind of audible noise but rather that it is saturated (DC?) and when you switch taps you hear some noise.

Easiest way to find out: Is the switching tap noise gone when you use your SE OTLs?
The A-Ts I ordered look pretty small in size.I cannot imagine that they can store a lot of energy under normal conditions (AC).

Wolfgang
Achim
Reply with quote  #25 
Wolfgang,

I purchased the ATs when I was running the SE OTLs - so I presume that my observation was made with the SEs. Sorry for the confusion. I don't think it's DC-related. I also do not think that my ATs would saturate from the tiny DC (<=40mV) that's on the Pinnacle outputs.

My ATs are in storage now. They are around 15kgs each, so might be not comparable with the ones you ordered.

Cheers,
Achiim






Kelvin Tyler
Reply with quote  #26 
Wolfgang,
In your post no.22 you mentioned the need for a CMOS switch. Did you mean a MOSFET switch?
Both the input gain and the comparator reference voltage would give control of the switching threshold in the scheme I had in mind. For fast, clean switching, I would probably use a Schmitt trigger comparator with a 311 or the like. The hysteresis need not be too big and would also be adjustable. The rectified signal would be fed to the comparator via some low-pass filtering which could be  adjustable as well. The 311 is an open collector device and is directly compatible with TTL logic.
I will sketch out more detail when I have time, and I have consolidated my ideas a bit more. Probably not until after Christmas now.
By the way, I have discovered some LCA715 devices which were left over from a previous project, so I will stick a couple  in the speaker feeds from my Pinnacles and see if I can detect their presence.
To be honest, I am very satisfied with the Pinnacle's performance as it is. I always try to keep well clear of clipping, especially with OTL amps. I actually have no need to implement this scheme, but being a bit of an experimentalist .......
Best seasonal wishes,
Kelvin
Wolfgang
Reply with quote  #27 
Kelvin,

my original idea was to use the LED meter as control unit and a simple CD4066, S1-S3, for switching. Like SW1/SSR1 is always 'on' if the power is turned on. Pure OTL is the default setting. The peak level, selected by threshold settings, would switch S2 & S3 to 'on' via, let's say BC 557 (T1) or something like that, S2 would switch SW2 & SW3 to 'on' and at the same time S3 would connect the control of S1 via resistor to ground so that its control is just below trigger level and open S1. As soon as the peak signal doesn't trigger T1, S2 &3 go to 'off' and S1 to 'on' again.

My A-Ts and a set of SSRs  have already been shipped and I hope we will soon have more clarity about certain things.

What about your SE 300Bs? They would  certainly benefit from such  a circuit.I am happy as can be with my version and still think they have the most outstanding bass and SQ and preserve the integrity of the music  no matter at which volume. I appreciate this mostly with classical music. But I have found out after testing  more powerful amps (just for kicks) that even with speakers over 100dB/1W/1m certain extreme dynamics are not 100% delivered by the SE 300Bs. It's not about clipping it's just about extreme dynamic jumps and peaks and speed. I would like to cure this if possible.

Happy Holidays,

Wolfgang
Wolfgang
Reply with quote  #28 
I hope everybody has survived x-mas.

For more clarity here is my idea of the basic circuit. CL would switch S2 and S3 on/off.4066_SSR.png 

Kelvin Tyler
Reply with quote  #29 
Wolfgang,
I have survived Christmas, just. Now have to contend with Hogmanay here in Caledonia!
Yes, that switching arrangement corresponds to your description except you have swapped the labels for S2 and S3. What I had in mind was to drive the MOSFET switches directly from the threshold logic via open collector inverting and non-inverting buffers. There are usually several ways to 'skin the cat' for this kind of thing.
However, I still do have strong reservations as to the overall viability of what you are suggesting. Your 300B DC SE is a current limited design with each tube idling at a hefty 125mA. Now imagine you have a level drive which just exceeds the chosen threshold. The circuit switches and, at this steady drive level,the load impedance is suddenly doubled, say. The result is an immediate halving of the output power. All transformers are constant power devices in principle, so this effect would hardly go unnoticed. To restore the power output the drive voltage would need to increase by root 2, and you would be back to where you were.
I suppose the point is an OTL may be able to give increased power into a higher impedance load, but requires more drive to do it.
I have tested these ideas via some simple tests with my Masterpiece, a scope and a selection of resistive loads. The MP has the same general topology as your 300B SE amps.
If what I am saying strikes you as nonsense, please say so.
Kelvin
Achim
Reply with quote  #30 
In my understanding the Autoformers serve as a kind of energy buffer through their magnetic reservoir. So when the amp runs out of steam and would normally clip, the iron adds that bit of needed current. That's also why there is this startup sequence where the magnetisation is first set up.

In my estimation and if my understanding is correct, this buffering is not compatible with dynamic switching. But Wolfgang will soon be able to tell us.


Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.