Many tweeters are rated at <<10W. Some might even be less than 2W to 3W, even though they are suitable, after a passive crossover filter, for systems approaching 100W.
For a steady state DC offset, the tweeter DC resistance, if no DC blocking capacitor is fitted, would be around 2/3rds to 3/4qrs of the rated impedance.
Now the VC is stationary and has less dissipation capability than a moving VC.
1.5Vdc into 3r0 is 3/4qrs of a Watt. Do you want that passing through your ultrafine VC wire?
What peak current flows prior to the opening of the protection circuitry?
I'm inclined to agree that around 1.5Vdc is a suitable detection value, but I would also use a DC blocking capacitor for the delicate VC in a tweeter, especially an expensive one or one that has had invested a lot of time trimming the crossover to best performance.
For a steady state DC offset, the tweeter DC resistance, if no DC blocking capacitor is fitted, would be around 2/3rds to 3/4qrs of the rated impedance.
Now the VC is stationary and has less dissipation capability than a moving VC.
1.5Vdc into 3r0 is 3/4qrs of a Watt. Do you want that passing through your ultrafine VC wire?
What peak current flows prior to the opening of the protection circuitry?
I'm inclined to agree that around 1.5Vdc is a suitable detection value, but I would also use a DC blocking capacitor for the delicate VC in a tweeter, especially an expensive one or one that has had invested a lot of time trimming the crossover to best performance.
Last edited:
When I say 1.5V, it is 1.5V peak. The protection will fire instant. No other noticeable delay than the one of your mechanical or solid state relay to open the circuit. For a mechanical relay, ~10ms. Lot faster for a MOS.
About your tweeter's fear, think about-it.
Or the amp is very powerfull, and this suppose your tweeter is behind a high pass filter, means isolated by a condenser and most of the time an attenuator. Or it is an active multyway, and, who think to feed a 1W tweeter with a 500W amp ?
I always try my speakers, whatever they are, with a 1.5V battery to ensure the + & - sens of mouvements. For thousand times longer that it will keep for the protection to fire. Never any problem.
Last, no one should admit a DC offset of 1.5V in the output of a working treble amplifier. Right ? And any amplifier will have a rail on 12V minimum, so 12V at the output in case of short circuit in a power device.
So i think it is OK that way.
Last edited:
It is a Scotish story: Jack le Scot came back from the town, a little drunk. His wife, very angry, asked him what remains of his money ( In the morning, he had 1£ in his pocket).
He said he still have half of it. She asks him how he spent the rest and he answer. "Mainly in women and wiskies".
iIt's coming. Rome was not built in a day.
We will use two of the four OPAs intead of one for comparing and amplifier the error signal, both to increase the bandwidth and to better separe 0 point adjustment (balance between input and output levels of the amps) and thereshold adjustement.
Last edited:
The choice in ampllifier protection circuits is driven by the OPAMP SERVO DECISION.
Amplifiers which make the commitment to support an Opamp DC Servo can economically support "high quality" protection circuits. The cost and size of the regulated +/- 15V power supply and relays(MOSFETS) for speaker and power protection dominates. High quality($3-$5) OPA627, OPA134/OPA2134 are selected for the DC Servo. Adding Esperado protection with one quad opamp(OPA4134) and one quad comparator(4LM339) offers high bang/buck value.
A "low performance" Esperado circuit is difficult to justify for "value".
Smaller PCBs with "basic functionality" need more publicity and review.
==============
Amplifiers which use a passive capacitor to ground for coarse DC control will naturally favor a protection circuit like the uPC1237 (with its 14 internal bjt's).
Amplifiers which make the commitment to support an Opamp DC Servo can economically support "high quality" protection circuits. The cost and size of the regulated +/- 15V power supply and relays(MOSFETS) for speaker and power protection dominates. High quality($3-$5) OPA627, OPA134/OPA2134 are selected for the DC Servo. Adding Esperado protection with one quad opamp(OPA4134) and one quad comparator(4LM339) offers high bang/buck value.
A "low performance" Esperado circuit is difficult to justify for "value".
Smaller PCBs with "basic functionality" need more publicity and review.
==============
Amplifiers which use a passive capacitor to ground for coarse DC control will naturally favor a protection circuit like the uPC1237 (with its 14 internal bjt's).
Here the simulation of the input stage in progress, full power, after rectifier stage, without and with an offset of 1.5 V in the amps.
First, for an amplfier of +-10V peak output, left with no offset, right with an offset of the amp of 1.5V.
Second, same thing with an amplifier of +-100V output.
The thereshold is here at its max limit for +-100V, sensibility can be increased by a factor of 10 for +-10V.
First, for an amplfier of +-10V peak output, left with no offset, right with an offset of the amp of 1.5V.
Second, same thing with an amplifier of +-100V output.
The thereshold is here at its max limit for +-100V, sensibility can be increased by a factor of 10 for +-10V.
Attachments
Last edited:
It is not "low performance". Just miniaturized and offering less options that not everybody needs. Just a microscopic , not expensive protection board, easy to be integrated in any amp.
Anyway, it is good to have the choice, don't you think ?
About servos, we could imagine to do a protection from-it. But consider it is not safe to use parts of an amp to protect-him (who knows from where a faillure can come) the second it would work only for protecting speakers from DC, and with the huge delay of the integration filter of the servo.
Last edited:
The following image shows where we are at the moment.
At 0s, power on. The protection (green trace) wait during 6,3s (time for the amplifier to settle in silence) before to connect the speakers.
As you can see (blue trace), at 10 seconds, a 1.5V offset was added at the out of the amplifier with a 1/10s duration.
The protection fires in a delay of less than 9/1000s and reconnect only after ~2s of delay after the end of the trouble .
Please, be patient for a couple of days to let-me verify everything.
For the components, we are at:
17 resistances
2 Adjustables
4 caps (+2 to bypass power lines near the OPA)
3 diodes.
1 quad OPA (Jfet inputs)
At 0s, power on. The protection (green trace) wait during 6,3s (time for the amplifier to settle in silence) before to connect the speakers.
As you can see (blue trace), at 10 seconds, a 1.5V offset was added at the out of the amplifier with a 1/10s duration.
The protection fires in a delay of less than 9/1000s and reconnect only after ~2s of delay after the end of the trouble .
Please, be patient for a couple of days to let-me verify everything.
For the components, we are at:
17 resistances
2 Adjustables
4 caps (+2 to bypass power lines near the OPA)
3 diodes.
1 quad OPA (Jfet inputs)
Attachments
Last edited:
Thank-you Andrew.
For those interested, or who can suggest improvements, the schematic (not approved for the moment).
Theory of operation:
Amps out are mixed by the two 10K while the amps in are mixed by the 2X47K.
The 50K adjustable tune the balance between the in and out inputs to null the difference as much as possible. It will be tuned between ~50K for a +and -100V peak to peak amplifier and around 10 Ohm for a +-10V.
Inputs are, for the moment normalized at 1V (peak to peak).
At the output of OPA.1, we have only the error signal. It set the threshold or sensitivity of the protection with the 10K adjust (10K for the 100V) Between 0 and 10K for 10V with a 1 to 10 gain in the next OPA.
The diode at its output rectify the negative errors. the next OPA invert the phase, in order to rectify the same way the positive errors. The 0.7V of the diodes provide a margin for the leakages, noises and balance errors as well.
The next OPA is in charge of the delays and fast switching.
The 470µF provide the "Power ON" delay of 6 seconds, while the 2MΩ and 2K resistances are in charge of setting a threshold in such a way that the protection will not fire with little voltage errors noises, and offset of the previous OPAs. On the other side, the diode the resistance and the 0.33µF cap allow a fast switching and a slow recovery (2s).
That's all for the moment. Comments and suggestions appreciated.
Attachments
Last edited:
I'm not fully satisfied with the input stage. It suffers from a drawback: The gain vary as you change the balance between input and output according to the various gain factors of the amps.
Not an real issue, as we can tune the gain factor of the second stage (Thereshold) but not snappy enough to think "got-it".
The solution shoud be to invert the two inputs, setting the "amp in" signals in the inverting OPA input. Because its impedance will not vary a lot as well as its levels.
But we need a high enough impedance in order to not add a charge to the preamplifier. This would lead to high values of the feedback resistance (~450K ?) if we want to keep a X10 gain. And their bad behavior with phases (parasitic capacitances of the inputs Jfet gates) wich are important when comparing high frequencies signals.
Ideas ? Suggestions ?
Not an real issue, as we can tune the gain factor of the second stage (Thereshold) but not snappy enough to think "got-it".
The solution shoud be to invert the two inputs, setting the "amp in" signals in the inverting OPA input. Because its impedance will not vary a lot as well as its levels.
But we need a high enough impedance in order to not add a charge to the preamplifier. This would lead to high values of the feedback resistance (~450K ?) if we want to keep a X10 gain. And their bad behavior with phases (parasitic capacitances of the inputs Jfet gates) wich are important when comparing high frequencies signals.
Ideas ? Suggestions ?
Last edited:
The new schematic does not show the dominance of power supply and MOSFET relays.
A small PCB with just the original protection plus soft-start is worth review.
(no soft stop, no fans, no input relays, no temp alarm)
ASSR-1611-301E solid state relay $6.00
LM339 quad comparator $0.50.
TL084C quad JFET opamp for control functions $0.50.
OPA4134 quad JFET opamp for DC servo + DC trigger $5.00.
So, $10 provides DC servo plus good protection.
A small PCB with just the original protection plus soft-start is worth review.
(no soft stop, no fans, no input relays, no temp alarm)
ASSR-1611-301E solid state relay $6.00
LM339 quad comparator $0.50.
TL084C quad JFET opamp for control functions $0.50.
OPA4134 quad JFET opamp for DC servo + DC trigger $5.00.
So, $10 provides DC servo plus good protection.
Thanks for your interest, LineSource.
As I said, it is better to have the choice. If you have a SMPS, you don't need any soft start, because it is supposed to be included in the PSU.
As the protection is able at its max sensitivity to fire with amps DC errors as low as ~30mV, go figure, silent stop will be done by the protection itself. And some will appreciate the very little size of the board that they can even wire directly in their amp at the output plugs.
In fact, it is this kind of protection and board that i Use in my amp since decades. I have a soft start included in my Main PSU, and, with the protection, never heard any parasitic or commutation noise during start, stop, or any commutation in the preamp.
I'm sure that lot of people were afraid by the apparent complexity of the first protection schematic. this one will demonstrate that it is not such a problem with the principle, as you pointed yourself.
I have done little changes in the schematic, according to my previous message, will publish-it soon after more testing.
Last edited:
Nice to see-you here, Steven.
I'm reaching the endline for this little circuit. Reversed the inputs in the comparator: Line input is better like this, because line levels have less range of sensibility than amps outputs. We accept, now, any sensitivity of the power amp up to 2Veff.
And any amplifier's power.
Sensitivity to the amp errors are up to ~30mV. Only vary of 10% at the max sensitivity that we can adjust by a 10 factor, depending of amp powers.
The two little caps in the inputs are really optional, depending of your amp's phase turns at 20Khz.
I'm thinking to design a little board that could use both SMD components and through holes ones, vertically positioned (old style) for the ones witch are afraid with SMDs. What do you think ?
I'm reaching the endline for this little circuit. Reversed the inputs in the comparator: Line input is better like this, because line levels have less range of sensibility than amps outputs. We accept, now, any sensitivity of the power amp up to 2Veff.
And any amplifier's power.
Sensitivity to the amp errors are up to ~30mV. Only vary of 10% at the max sensitivity that we can adjust by a 10 factor, depending of amp powers.
The two little caps in the inputs are really optional, depending of your amp's phase turns at 20Khz.
I'm thinking to design a little board that could use both SMD components and through holes ones, vertically positioned (old style) for the ones witch are afraid with SMDs. What do you think ?
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.