| Sorry Nelson, but I have found an error ... - Click HERE for Original Thread |
| toshiba |
... in the Aleph ONO schematics drawn by Wayne.
Correcting this error should solve all problems concerning the Ono mute circuit in cloned PCB's. Perhaps other problems too ...
Q25, labeled as "9610" in the original schematics of Aleph Ono service manual is wrong.
First it has to be a "610" and second, Drain and Source have to be swapped.
Only then, Q25 works as a simple voltage regulator with an output rail of about 24V to drive a 24 Volt-Relay.
Nelson, or do you work with 40V-Relays ??? :Pinoc:
The rest of the mute circuit is ok. The need of any additional components like shunt resistors (pquadrats PCB's) isn't required.
Best regards
toshiba with small "t"
P.S.
Nelson I'm sad, that you haven't answered to my quote |
|
|
| Fred Dieckmann |
The circuit looks fine to me........
From page three of the ALEPH ONO SERVICE MANUAL
REV 0, 1, 2, 3 9/30/1997 :
"Q25 and Q26 are used to power the mute relays."
How it appears to work to me:
The Mosfet act as switches they are low impedance when the gate to source voltage is greater than about 4 volts. This circuit is to short the outputs to ground until the supply comes up and the circuit stabilizes. R77 (750K) and C40 (220uF) provide a delay as the cap charges up and Q26 turns on. Q25 is already on when the 40 volt rail comes up. D11 clamps inductive voltage spikes across the relay coil and D10 provides hysteresis so the relay switches cleanly and does chatter during the switching.
The gate to source voltage for Q25 is about 12V after the 40 V rail comes up. You want to output to mute quickly when the power goes off and Q25 will go high impedance quickly while Q25 will still be swicthed on. You want to unmute after a delay when powering on and to mute quickly when powering down. 48 relays are very common for telecom use and should work for this circuit.
See anything wrong with the audio circuit you want to tell us about? |
|
|
| UrSv |
I can only agree with Fred as it seems fairly straight-forward and the description Fred gave matches my understanding of the circuit for the Aleph P which I am currently building and which is the same and which in that case also would be faulty. Funny nobody has mentioned it since so many built these? Even more funny that I remember threads about it as well as reading the service manual where it is explained.
Personally, if I did not understand a circuit at all and wanted to I would probably ask how it works first and then claim it is incorrect...but that is just me. |
|
|
| faustian bargin |
| quote: | Originally posted by UrSv
...but that is just me. |
i'm guessing it's not just you...probably most people.
claiming certainty in almost anything is just asking for a big serving of foot...especially when one is in the company others who probably know more. |
|
|
| Duo |
Man oh man.
Expert: X being the unkown value; Spert being a drip under pressure.
Mr. Pass knows what he's doing. How else would his work be so respected in the audio community?
imho: I personally think his work should be commended. It's an excellent contribution to this forum and its people for sure. :) |
|
|
| jh6you |
| quote: | Originally posted by Fred Dieckmann
Uh Oh another expert.........
|
I like your technical response very much.
But, the title is totally wrong.
JH |
|
|
| grataku |
| quote: | Originally posted by toshiba
... in the Aleph ONO schematics drawn by Wayne.
Correcting this error should solve all problems concerning the Ono mute circuit in cloned PCB's. Perhaps other problems too ...
Q25, labeled as "9610" in the original schematics of Aleph Ono service manual is wrong.
First it has to be a "610" and second, Drain and Source have to be swapped.
Only then, Q25 works as a simple voltage regulator with an output rail of about 24V to drive a 24 Volt-Relay.
Nelson, or do you work with 40V-Relays ??? :Pinoc:
The rest of the mute circuit is ok. The need of any additional components like shunt resistors (pquadrats PCB's) isn't required.
Best regards
toshiba with small "t"
P.S.
Nelson I'm sad, that you haven't answered to my quote |
They should put this post in the dictionary as the meaning for the word _dork_
I am sure looking forward for your next 'critique' of Nelson's work since I am about to build an aleph :rolleyes: |
|
|
| janneman |
| quote: | Originally posted by Fred Dieckmann
The circuit looks fine to me........
From page three of the ALEPH ONO SERVICE MANUAL
REV 0, 1, 2, 3 9/30/1997 :
"Q25 and Q26 are used to power the mute relays."
How it appears to work to me:
The Mosfet act as switches they are low impedance when the gate to source voltage is greater than about 4 volts. This circuit is to short the outputs to ground until the supply comes up and the circuit stabilizes. R77 (750K) and C40 (220uF) provide a delay as the cap charges up and Q26 turns on. Q25 is already on when the 40 volt rail comes up. D11 clamps inductive voltage spikes across the relay coil and D10 provides hysteresis so the relay switches cleanly and does chatter during the switching.
The gate to source voltage for Q25 is about 12V after the 40 V rail comes up. You want to output to mute quickly when the power goes off and Q25 will go high impedance quickly while Q25 will still be swicthed on. You want to unmute after a delay when powering on and to mute quickly when powering down. 48 relays are very common for telecom use and should work for this circuit.
See anything wrong with the audio circuit you want to tell us about? |
Fred,
Sorry, but your explanation still doesn't help me. Most of us know that we need to mute quickly and unmute fast. Why the need for two mosfets, complementary at that, when one could do the trick? toshiba might have been wrong, but nobody yet has proven that or even argued against him. I think it isa a bit cheap to say "...Mr Pass knows what he does...". Hell, we ALL know what we do, don't we? Let's try to keep this discussion intelligent.
Jan Didden |
|
|
| paulb |
| quote: | Originally posted by janneman
Why the need for two mosfets, complementary at that, when one could do the trick? |
I suspect it has to do with PassLabs having a lot of leftover FETs that failed the hand-selection for matching.
When someone like toshiba comes out with an arrogant-sounding thread title and even refers back to his own earlier post pointing out another supposed error ("I'm sad you haven't answered..."), they can count on a few flames. |
|
|
| UrSv |
| quote: | Originally posted by janneman
Fred,
Sorry, but your explanation still doesn't help me. Most of us know that we need to mute quickly and unmute fast. Why the need for two mosfets, complementary at that, when one could do the trick? toshiba might have been wrong, but nobody yet has proven that or even argued against him. I think it isa a bit cheap to say "...Mr Pass knows what he does...". Hell, we ALL know what we do, don't we? Let's try to keep this discussion intelligent.
Jan Didden |
Reading a bit more carefully I see that Fred possibly made a small mistake in his explanation. The sentence:
"You want to output to mute quickly when the power goes off and Q25 will go high impedance quickly while Q25 will still be swicthed on. You want to unmute after a delay when powering on and to mute quickly when powering down."
Should probably read:
"You want to output to mute quickly when the power goes off and Q25 will go high impedance quickly while Q26 will still be swicthed on. You want to unmute after a delay when powering on and to mute quickly when powering down."
So essentially the circuit switches Q25 on quickly at power up (since it's gate voltage is presented at the same instant) and Q26 will be switched on when C40 is charged to the right level (it's gate voltage is presented with the delay of the charging of C40) and thus both FETs are switched on and the relay is activated. When power is switched off then Q25 will switch off quickly (as the gate voltage disappears at the same time) and the relay will be deactivated. Q26 still remains switched on (as C40 is still charged) but since Q25 is switched off then the relay can not be activated. This provides slow muting off at power on and quick muting on at power off.
Could you explain how to do it with just one FET? |
|
|
| Fred Dieckmann |
paulb beat me to the punch. The circuit is neither overcomplicated
nor expensive. It appears to me to be a clever and reliable way to do it using parts that he is buying in large quantities. (read volume discount) Telecom relays are very inexpensive and very reliable also.
"Could you explain how to do it with just one FET?"
I kind of like the two fet circuit. I'll bet you can tell us how to do it since I think you understand the circuit as well as I do. Go for it. I need to go get ready to go to the doctor about my blood pressure (again..... it's getting to be a real pain)
I came not to flame but merely to try to shed some light on the circuit. "See anything wrong with the audio circuit you want to tell us about?" was gentle barb comparded to some I have handed out this week. Somehow I think Mr. Pass was probably amused and not offended by toshibas post. I seem to remember Mr. Pass
getting second guessed on audio circuits several times without getting ruffled. He is a gentleman and a scholar while I am a trouble maker and provocateur. Many of my post seem to arose the response one would expect from one who's car had just been keyed.
"What do you know? See an ordinary person spend his life avoiding tense situations. Repoman spends his life getting into tense situations." |
|
|
| tortello |
Hello,
here is a quick simulation of the timer's behaviour.
The top ("ps", blue) is the power supply trace, about 40V when I power-on the preamplifier at 0 seconds, then about 0V after 20 sec., when I decide to turn-off the power switch.
The "output" (red) trace is the relay contacts' trace, monitored with a pull-up resistor.
The other two traces are the relay coil's pins.
There is a little oscillation, I don't know if it is a parasitic phenomenon, or maybe an effect due to the rushed simulation:
these graphs are only to see the operating behaviour of the circuit.
Marcello |
|
|
| x-pro |
I would not agree with toshiba - the circuit would work fine just as Fred described, thought with some overvoltage on the relay (it will see about 35-36 V - most 24V relays will survive it. I personally would add a resistor in series with the relay to drop the voltage down to 24V.
However I would be a bit critical from another point of view. I did use a very similar circuit in a production amplifier and found few years later that relays did not really like "soft switching" where the relay voltage rises relatively slowly, as FET just amplifies the input voltage. As a result of it the relay contacts did deteriorate somewhat over the time. So later I did switch to a Schmitt-trigger kind of circuit shown on the picture (sorry for a poor quality, as I only could use USB camera to get it in here). We didn't have any relay problems since.
Al |
|
|
| janneman |
| quote: | Originally posted by UrSv
[snip]
Could you explain how to do it with just one FET? |
A diode across R77, pointing to the left?
Jan Didden |
|
|
| jh6you |
| quote: | Originally posted by Fred Dieckmann
See an ordinary person spend his life avoiding tense situations. |
But, not by giving others tense... ;)
JH |
|
|
| tortello |
| quote: | Originally posted by x-pro
... thought with some overvoltage on the relay (it will see about 35-36 V - most 24V relays will survive it. I personally would add a resistor in series with the relay to drop the voltage down to 24V.
Al |
I think Nelson uses 48V relays: even with only 33-34Vdc of pull-in voltage, they seem to work well.
Marcello |
|
|
| Fred Dieckmann |
| 48 volt relays are as common as dirt in the USA due to their wide use in telecom. You can get them pretty cheap since they are manufatured for hign volume use in telecom. We used about 12 million a year on one type of line card that I worked on a few years ago. The assembly plants for many Japanese relays are in China to reduce labor cost. They will operate at lower voltage than 48 since the is the nominal value for talk battery and it's range goes to a lower voltage to allow for battery discharge. You don't need series resistors in you use the right relay for the right supply voltage the coil resistance is fairly high to keep power dissapation low. I would be be willing to bet that Passlabs cound build the entire circuit for less what you our I could by one 24 volt relay in small quantities. I could probably do a cost model of what they spend for the circuit to within a 10% to 20% window. I guess designing a relay circuit for relability and cost is harder than most people thought. By the way what high volume electronic devices use 24 volts? Cars use 12 V, most logic circuits use 5 V, and telecom uses 48 V, I don't think I have ever seen a 24 volt relay although I am sure they make them. I'll bet you you would pay more for them than 48 V relays. |
|
|
| moamps |
| quote: | Originally posted by janneman
A diode across R77, pointing to the left?
Jan Didden |
Hi,
not fast enough, PS 40 V slowly drop down to zero and slowly discharge cap thru your diode.
Regards |
|
|
| janneman |
| quote: | Originally posted by moamps
Hi,
not fast enough, PS 40 V slowly drop down to zero and slowly discharge cap thru your diode.
Regards |
Well, if it is fast enough for the top fet, it's fast enough for the diode.
Jan Didden |
|
|
| janneman |
Sorry Fred, no intention to flame or anything. Just got exited seeing someones well-intended and well reasoned post brushed aside.
Your explanation DOES make sense, of course, as (most) all of your posts do.
Jan Didden |
|
|
| moamps |
| quote: | Originally posted by janneman
Well, if it is fast enough for the top fet, it's fast enough for the diode.
|
Hi,
This isn't same. R78 and 74 and LED(?) make voltage divider (factor k) for sensing dropout of PS. Upper FET will very sharp switch off when PS drop below Vgs*k. With this voltage divider you can adjust that upper FET switch off at 30-35V, what is with your diode impossible.
Regards |
|
|
| Fred Dieckmann |
I think you might want to mute before supply discharged enough to discharge the cap enough to turn off the MOSFET. That is a slow transition also and they relay will chatter trough this transition without hysteresis.
Something this looks like the type of relay I think they might by using:
http://oeiwcsnts1.omron.com/pdfcata...2568FF0050EF69/$FILE/D20G5V20100.pdf |
|
|
| x-pro |
Fred is right, you'll need this second MOSFET or a similar trick if you power the relay from the main supply rail. I didn't mention that for my circuit I've used a separate supply with a small smoothing capacitor, so on switch-off this supply goes down very quickly.
Al |
|
|
| Fred Dieckmann |
| 24 volts..... I stand corrected by the gracious lady with all deference due. I am I right in assuming its a 2 form c? |
|
|
| Fred Dieckmann |
I will be more careful since I am being monitored. I shudder to think of the reaction to my interpretation of how the rest of the circuit works..........
There is only one thing in the world worse than being talked about, and that is not being talked about. - Oscar Wilde |
|
|
| janneman |
| quote: | Originally posted by moamps
Hi,
This isn't same. R78 and 74 and LED(?) make voltage divider (factor k) for sensing dropout of PS. Upper FET will very sharp switch off when PS drop below Vgs*k. With this voltage divider you can adjust that upper FET switch off at 30-35V, what is with your diode impossible.
Regards |
Indeed. I stand corrected.
Jan Didden |
|
|
| UrSv |
| quote: | Originally posted by Fred Dieckmann
24 volts..... I stand corrected by the gracious lady with all deference due. I am I right in assuming its a 2 form c? |
And I think the 24 V relays are quite common as I could not remember ever seing a 48 V one in my catalogues. For instance http://www.elfa.se/ seem to have some 80 different relays using 24 VDC but only 3 that use 48 VDC (which I did not know about). As a matter of fact I thought 5, 12 and 24 was standard issue and 48 V was the exception in the everyman supply chain. |
|
|
| jam |
Nice one Tazzy!
It is not nice to question the master. |
|
|
| Dane |
Schematic does mention "Right Channel Shown",
So there is a Left Channel, and another Mute Relay "K2".
The 2-24V coils are probably in series as in the Aleph Preamps. |
|
|
| swede |
Just a thought... my two pennies... on mixing analogue and digital techniques for the muting...
I've got quite a bit more caps on my PS and on my Aleph P boards, than used in the original schematics. Just because it was easier for me, since I had some leftover in my scrap box.
This makes the time for the "turn on thump" to be around 15 - 20 seconds...
The muting circuit goes on after only one - two seconds, which IS enough time, since I need less than 100ms to boot up my PIC (getting the power to the right level + restart of the chip). The first thing it does is to reduce the volume to 0, then go to sleep for 20 seconds and then set the volume to the "last" value.
Just my two...
The muting (analogue part) works fine, and is based on the Aleph P 1.7 schematic, althoug broken off on a separate board.
//magnus |
|
|
| Nelson Pass |
I worked that circuit out about 1975 (with Bipolars originally)
and have been using it since.
:cool: |
|
|
| toshiba |
| quote: | Originally posted by Karen
... PICTURE OF RELAY |
Thanks a lot Karen !!
Please can you deliver Mr. Pass the following message : "Sorry for attacking
and chosen a strong title like this one for explaining what I have changed. I'm someone
who often likes to do something in a disrespectfully way."
As I wrote the thread, I thought, that the experts are coming up, analyze my changes
and kill them with words like : "You crazy crackpot from germany ... what you have done
can't work ...."
I didn't expect that anyone replies with words like : "Ok you are right and it's your way to do so."
But what happened ? The experts explain me the function of the original mute circuit and the
non experts announce their ordinary comments.
Sometimes it's better to find his own way.
toshiba
P.S.
I really respect the work of Mr.Pass. His amps sounds incredible.
Sure, he is one of the greatest, perhaps the greatest !
But could it be the reason, that it is forbidden to do some changes ? |
|
|
| Fred Dieckmann |
"Nice one Tazzy!"
You are allowed ( and probably encouraged by some) to doubt me. You will find that is an international sport. It keeps me on my toes and gives me a chance to start fights over several continents on any given subject. If Steve Eddy gets involved, they can go on for days.......... |
|
|
| Fred Dieckmann |
"48 V was the exception in the everyman supply chain."
I guess I was in Telecom too long. I have seen thousands and had about 6 six relay maufacturers on my approved vendors list. |
|
|
| jh6you |
| quote: | Originally posted by Fred Dieckmann
"Nice one Tazzy!"
You are allowed ( and probably encouraged by some) to doubt me. |
Nelson Pass = Fred Dieckmann ...?
:bigeyes:
JH |
|
|
| pinkmouse |
| quote: | Originally posted by jh6you
Nelson Pass = Fred Dieckmann ...?
|
Now that sock puppet got right past us mods, tee hee;) |
|
|
| davesaudio |
I seem to remember telecom equipment at 24v, more than one C.O with such plant
and military vehicles /hardware? |
|
|
| jh6you |
Question to Fred Dieckmann.
As you explained, Q25 (9610) is introduced to support a quick mute.
When the rail voltage drop curve passes 32V, the relay will die.
(Pickup voltage of the relay: 18V). Do you agree?
JH |
|
|
| UrSv |
| quote: | Originally posted by toshiba
Thanks a lot Karen !!
Please can you deliver Mr. Pass the following message : "Sorry for attacking
and chosen a strong title like this one for explaining what I have changed. I'm someone
who often likes to do something in a disrespectfully way."
As I wrote the thread, I thought, that the experts are coming up, analyze my changes
and kill them with words like : "You crazy crackpot from germany ... what you have done
can't work ...."
I didn't expect that anyone replies with words like : "Ok you are right and it's your way to do so."
But what happened ? The experts explain me the function of the original mute circuit and the
non experts announce their ordinary comments.
Sometimes it's better to find his own way.
toshiba
P.S.
I really respect the work of Mr.Pass. His amps sounds incredible.
Sure, he is one of the greatest, perhaps the greatest !
But could it be the reason, that it is forbidden to do some changes ? |
Oooh, so that is what you were doing. Explaining what you had changed. I interpreted "Correcting this error should solve all problems concerning the Ono mute circuit in cloned PCB's. Perhaps other problems too ..." as if there was an error or even a problem in the circuit and that it would not work.
You are more than welcome to change anything you wish...at least I have nothing against that. I just think very few people understood your post as telling us about the changes you had made as it sounded like something else. |
|
|
| moamps |
Hi all,
I make some experiments with this delay circuit, and results are interesting for me.
First, in service manual for preamp I don't see two delay circuit (only two relays) and my conclusion is that two relays are connected in series.
Second, I measure several 24V relays and resulting voltage are Von about 18V, Voff about 7V.
Third, Q25 (in my experiment 9630) start conduct when Vgs is about 3V, and this happened (if D is LED with Va-k-on about 2V) when Vcc is greater then 12V with value of R78 and R74 referred to service manual.
Where is fault? Q25 will switch off when Vcc drop at 12V. In this moment relays are already deenergised (or in same time) because Voff relays are about 7V (two in series, 14V). I try shortening Q25, and Voff (and time off) is same with and without. With this value of R78 and R74, Q25 is useless.:scratch:
I connect 1k resistor parallel to R78, then Voff of Q25 rise to about 24V, and then this circuit work fine.
Maybe elements in original design have slightly differ characteristics, but my theoretically and practically point of view is that this circuit need some revision. :scratch:
Regards
test circuit |
|
|
| Fred Dieckmann |
"Nelson Pass = Fred Dieckmann"
"Now that sock puppet got right past us mods, tee hee"
Don't even make a joke like that, I already get trashed for using the name "Norm Thagard" the name of a very famous astronaut and audio amp hobbiest, who is friend of Mr. Pass I believe. It was such a obvious joke that I though everyone would know after one post. I believe Grey didn't and was extremely displeased. |
|
|
| moamps |
| quote: | Originally posted by Fred Dieckmann
[BWell don't leave just hanging. Nobody else can figure it out either.
Straighten us out please!
out. [/B] |
Hi,
OK, let's start from the scratch:
D? is a regular LED with Vak = 2.5 V. There are two relays in the circuit with nominal operating voltage Vr = 24V. (the original service manual says "Q25 and Q26 are used to power the mute relays. The front panel LED must function for Q25 to turn on and pull the mute relays off").
Bottom MOSFET is irrelevant because it functions as a short circuit when the circuit is powered off.
The circuit operates as follows: When the Vunreg starts dropping, relays voltage starts dropping as well until Vunreg drops below Voff=K*Vgs(th)+Vled, where K =(R78+R74)/R78, Vgs(th) = ca. 3V and Vled= ca. 2.5V. Then the MOSFET switches off relatively quickly (transfer function knee), which switches off the relays as well. In this case Voff = ca. 12V. However, by the time Vunreg drops below Voff=12V, the relays have already reached Vdrop-out value and deenergized. Switching off relays via MOSFET Q25 is therefore useless because they are already switched off, i.e. Q25 can be left out.
If we want to increase Voff(MOSFET), we need a lower value of R78 or a higher value of R74 or to increase Vled (by placing more LEDs in series or replacing them with a Zener diode).
In this case, if we want Q25 to switch off the relays when Vunreg drops below Voff = 30V, keeping 1 LED in the circuit and the value of R74=3.32k, we can calculate the value of R78 using the following formula: R78=Vgs(th)*R74/(Voff-Vd-Vgs), R78=3*3.32/(30-2,5-3)=ca 400ohm. If we take the standard value of R78=470ohm, we will get Voff=ca 27V. It would be possible to use a higher value of Voff but, since I don't know what's the value of Vripple on Vunreg, Voff=30V seems to be the best choice here.
If anyone finds a flaw in my method, I'd appreciate their comments and feedback. Also, I apologize if my post is too long and/or too boring.
Regards |
|
|
| jh6you |
When the relay activates, Vgs of Q25 is 12V and coil voltage is 24V.
When the power turns off, the relay coil acts as an inductor.
Then, the relay coil induces emf, trying to maintain the original coil voltage of 24V until Vgs of Q25 reduces down to 4V. This might happen in short time, and sudden off of Q25.
Q26 turns on the relay, the relay turns off Q25, and whole system goes off.
Am I really silly?
JH |
|
|
| TheMyxin |
[QUOTE]Originally posted by moamps
[B]Hi all,
I make some experiments with this delay circuit, and results are interesting for me.
First, in service manual for preamp I don't see two delay circuit (only two relays) and my conclusion is that two relays are connected in series.
It depends if you use two or one PCB-Supplyboard. Some DIY-Ono-Clones use only one PCB-Board and in this case you connect two Delay relais from one
40 Volt Supply rail from the regulation part. In my case i use only one PCB-Supply-Board. I think in the original
Xono-Design there are two PCB-Supply boards right ?
For practical reason (its much easier to get 24Volt Relais instead a
48Volt relais) PQuadrat (who made the
PCBs) and i for example are using the meder relais 24volt (Part-Number DIL24-2C90-63L). The coil resistens in the data-spec is 2 KOhm.
On the PCBs there is a pre-resistor R19.
The Supply-Voltage for the Relais i have measured in my case 42V without loads.
With R19=1.5 Kohm i had switching
problems. So it would be there nice if you could check that with this pre-resistor R19 and two Relais from one PCB-Supply and one regulation.
Thanks a lot.
Is in this case a Resistor R19= 3,3 Kohm
i observed no swichting problems since them. But it could be in my case a problem with a faulty R77 and/or Q25
(see my other thread swichting problem
aleph ono) |
|
|
| moamps |
| quote: | Originally posted by TheMyxin
[QUOTE]Originally posted by moamps
[B]
So it would be there nice if you could check that with this pre-resistor R19 and two Relais from one PCB-Supply and one regulation.
|
Hi,
I'm not sure how look your pcb, and where you place R19. I made for you quick test and circuit below work just fine. If your schematic is differ, please put it here. My suggestion (revision) is changing R78 to 470ohm (q25 will switch off about 27V).
If you are trying this circuit before correcting error with direction of D11 (earlier posts), maybe is your Q25 burned. R77 can't be burned.
Regards |
|
|
| moamps |
| quote: | Originally posted by jh6you
When the relay activates, Vgs of Q25 is 12V and coil voltage is 24V.
When the power turns off, the relay coil acts as an inductor.
Then, the relay coil induces emf, trying to maintain the original coil voltage of 24V until Vgs of Q25 reduces down to 4V. This might happen in short time, and sudden off of Q25.
Q26 turns on the relay, the relay turns off Q25, and whole system goes off.
Am I really silly?
JH |
Hi,
you aren't silly.
Acting relays like inductor is not important here. Back EMF is supressed with D11. Look my measurement at circuit one post up.
Regards |
|
|
| Fred Dieckmann |
Go look at post number two in this thread. This is a relay driver with two transistors. How hard can it be to understand? One 24 V relay will handle the voltage the coil will see. Two of them would be in parallel if the schematic is correct. It shows one only channel and no coils in series are shown.
"Vunreg drops below Voff=12V, the relays have already reached Vdrop-out value and deenergized."
Gee..... the data sheet said the drop out voltage for the 24 volt relay is a minimum of 2.4 Volts. The voltage for a blue LED is 3.5 to 4 volts I believe.
http://ctlgserv.mew.co.jp/acg/cgi/m...link.cgi?l=ds2y
I was being very sacastic about being straightened out on how it works. The fact that I think I explained it on the second post of the thread and posted the data sheets might have been a clue. The next time you rocket scientist want to deconstruct a simple circuit, at least READ THE DATA SHEETS. Somehow I think Mr. Pass is smart enough to design a relay driver and careful enough to draw the schematic correctly. This whole thread is one of the silliest things I have ever read. |
|
|
| moe29 |
"This whole thread is one of the silliest things I have ever read."
Agreed. |
|
|
| Fred Dieckmann |
| Get rid of the 1 K resistor. The turn on voltage is about 17 volts max for the relay Passlabs specs. I don't think Nelson is the one adding extra parts. |
|
|
| moamps |
| quote: | Originally posted by Fred Dieckmann
I was being very sacastic about being straightened out on how it works....... |
OK, Dickmann,
so I didn't get your sarcasm. I'm just a B.Sc.E.E. and I don't get it when people get all metaphorical or whatnot on me (in a foreign language too) when I'm trying to discuss some serious issues. Reading your posts, it was only clear to me that you had no idea on how the thing works in practice (you also used to claim relays were 48V, remember?) so I truly thought you needed some 'straightening out'. Now I don't know why I even bothered as you can obviously think of no other argument besides Mr. Pass-knows-best mantra (with all due respect to Mr. Pass) that you keep repeating post after post.
"Gee..... the data sheet said the drop out voltage for the 24 volt relay is a minimum of 2.4 Volts. The voltage for a blue LED is 3.5 to 4 volts I believe."
Now, you could also benefit from reading data sheets more carefully as this one clearly says that the MINIMUM is 2.4V. I don't see it anywhere written what's the maximum. Do you? I'm telling you that IN PRACTICE Vdrop-out is about 7V for a 24V relay; blue LED doesn't change anything. Last time I checked, my instruments were working perfectly well.
Unless you can prove me wrong, I'll keep claiming that there is a flaw in the circuit design - Q25 switches off too late.
Truly yours,
Rocket Scientist |
|
|
| Fred Dieckmann |
I will be glad to talk to you after you read the data sheet. It's a little hard to discuss theory when you don't have the data. You can write equations you want but plugging values for the variables in off the top of your head won't tell you much.
12V -3V is 9V which is a hell of lot more than 2.4 volts the top fet will mute the relays long before the 2.4 volt drop out voltage. I can hardly wait till will audio circuit.......... |
|
|
| TheMyxin |
| quote: | Originally posted by moamps
Hi,
I'm not sure how look your pcb, and where you place R19.
I will post the Thread where you can see the photo of the Alpeh Ono Boards from the top. R19 is the resistor between the Relais and the MOSFET on the Ono-Boards
http://www.diyaudio.com/forums/show...5&highlight=Ono
The PCB-Supply is on the right PCB and
feeds direclty R19 and the Relais.
As i understand correctly R19 is in the cicuit-scheme directly behind D10 and before the coil of the relais. So the
coil restors and the pre-resistor r19 build a voltage dividivier for the Relais.
correcting error with direction of D11 (earlier posts), maybe is your Q25 burned. R77 can't be burned.
No i have correct the Direction of D11 but perhaps Q25 was really damaged
because i have made my first tests
without the correction. So know i have
a new R77 and new Q25. R77 was at
the start only 0.25W and now has 0.6W
(i was afraid that 0.25W is to low !)
Regards |
|
|
|
| Nelson Pass |
Referring to the schematic by Fred on the first page of this thread, there are a couple of comments which might shed some additional light on the circuit.
The diode from Gate to Drain of Q26 is there to prevent the Gate of Q26 from getting more positive that its rating, otherwise ultimately the Gate would be at the + rail value. When Q26 conducts, its Drain goes toward ground until it hits about +3 V,
at which point the current starts flowing through the diode.
Often this circuit has been implemented differently around Q25, where a supply regulator is being used. The Source of Q25 is attached to the positive unregulated supply, and the Gate is attached (through a resistor) to the regulated supply. When the difference between unregulated and regulated supply becomes less that the Vgs of Q25, the relay is shut down. This makes an excellent sensor which turns the relay off when there is no longer enough unregulated supply to hold regulation, the "off" condition corresponding nicely to this. |
|
|
| moamps |
| quote: | Originally posted by TheMyxin
|
Hi,
R77 is 750K resistor, current thru them can be in worst case I=V/R=40V/750000=ca 50uA. Power dissipation on R77 is then P=V*I= 2mW !!!!! , 0,25W is more than enough. Preresistor R19 can be between zero (maximum heating of relays, Q25, Q26) and about 1k (from my experiments).
Regards |
|
|
| Fred Dieckmann |
I guess that means that it works as drawn....... imagine that! I suppose the threshold for Q25 turning off could be very easily set for a higher supply voltage with a smaller value for R78. That that seems to be a concern for a couple of people....... Also, I guess if one did not have as much supply voltage to play with that a 9 volt zener clamp gate to ground for Q26 instead of the gate to drain diode could get you another 3 volts to play with across the relay coil.
I think it is a very nice circuit for using left over mosfets from matching. Adapting for bipolars should be very easy too. Want not waste not. I can't understand the confusion over it once one knows the relay threshold voltages. |
|
|
| moamps |
| quote: | Originally posted by moamps
.
Maybe elements in original design have slightly differ characteristics, but my theoretically and practically point of view is that this circuit need some revision. |
| quote: | Originally posted by Nelson Pass
.
Often this circuit has been implemented differently around Q25, where a supply regulator is being used. |
WhoDaMan ! :up:
Thank you Mr. Pass
Best regards |
|
|
| thylantyr |
I need some help understanding the original schematic.
If my power supply is 40V peak, when listening
to music, the rails drop to rms value, approx 28v
under heavy load, ie, the rails bounce from 28v - 40v.
Does the circuit mute when the unregulated voltage
drops below 28v ?
I was looking at the Rob Elliot "fast" mute circuit, pretty cool
circuit, but more complex. I'm trying to figure out which
one to use :bigeyes: |
|
|
| moamps |
| quote: | Originally posted by moamps
If we want to increase Voff(MOSFET), we need a lower value of R78 or a higher value of R74 or to increase Vled (by placing more LEDs in series or replacing them with a Zener diode). |
| quote: | Originally posted by Fred Dieckmann
I guess that means that it works as drawn....... imagine that! I suppose the threshold for Q25 turning off could be very easily set for a higher supply voltage with a smaller value for R78. That that seems to be a concern for a couple of people....... |
"Friends, Romans, Countrymen, lend me your ears..." |
|
|
| Jocko Homo |
Let me see if I have this right.........
Nelson designs a circuit 28 years ago......puts it in countless units in the field........and now we have experts claiming it is wrong.
Don't you think that he would have lots of angry customers sending them back by now, and that he would have known about it?
Good grief, Charlie Brown.
Jocko |
|
|
| Fred Dieckmann |
"If my power supply is 40V peak, when listening to music, the rails drop to rms value, approx 28v under heavy load, ie, the rails bounce from 28v - 40v."
It's 40 volts DC and it is a preamp! I don't think your supply will drop to 28 volts except when you turn the supply off.
Sorry Moamp I missed your threshold remark in the midst of you telling us the circuit didn't need the upper mosfet. It is funny that the circuit didn't work when I explained it and the same circuit magically works when Mr. Pass logs on to say it works.........
Et tu moamp? |
|
|
| jh6you |
I know you enjoy the silly things. Ha!
JH |
|
|
| thylantyr |
it is a preamp!
I see said the blind man.
From what I gather on this website,
not many people like speaker relays
on power amps. hehe ...... |
|
|
| moamps |
"Nelson designs a circuit 28 years ago......puts it in countless units in the field........and now we have experts claiming it is wrong."
"It is funny that the circuit didn't work when I explained it and the same circuit magically works when Mr. Pass logs on to say it works........."
You must be dislexic or something because I certainly have never written the thing doesn't work at all but that it NEEDS REVISION because the design we have been discussing here causes people some problems, namely things around Q25. Then Mr Pass comes along and says:
"Often this circuit has been implemented differently around Q25, where a supply regulator is being used."
Can't you read?! For God's sake, it's written in YOUR MOTHER TONGUE by Mr Pass himself.
Eppure si muove, people, eppure si muove... |
|
|
| Jocko Homo |
"Sorry, Nelson. but I have found an error......"
So, tell me, where is the error?
Different implementation is not an error.
Error means it does not work.
Maybe it should have read:
"Nelson, I have no idea how this works. Am I missing something?"
Jocko |
|
|
| jh6you |
| quote: | Originally posted by moamps
Eppure si muove... |
| quote: | Originally posted by Jocko Homo
mah ah-thor-ah-tah!" |
Am I an inhabitant of Mars...?
JH |
|
|
| UrSv |
| quote: | Originally posted by moamps
--snip--
"Often this circuit has been implemented differently around Q25, where a supply regulator is being used."
Can't you read?! For God's sake, it's written in YOUR MOTHER TONGUE by Mr Pass himself.
Eppure si muove, people, eppure si muove... |
It is not my mother tongue obviously but I read it exactly as I assume Fred and Jocko (clearly by his statement I feel) did. That is not the way you read it as I read it has been built differently and nothing else. Not that I follow the Ono discussion very closely but who are all these people having problems? |
|
|
| stefanobilliani |
| quote: | Originally posted by moamps
&Eppure si muove, people, eppure si muove... |
...Malgrado l'innerzia imposta...:) |
|
|
| grataku |
| quote: | Originally posted by stefanobilliani
...Malgrado l'innerzia imposta...:) |
stefanuzzo,
inerzia si scrive con una -n-. |
|
|
| grataku |
| quote: | Originally posted by Jocko Homo
"Sorry, Nelson. but I have found an error......"
So, tell me, where is the error?
Different implementation is not an error.
Error means it does not work.
Maybe it should have read:
"Nelson, I have no idea how this works. Am I missing something?"
Jocko |
Jocko I think you maybe confusing MoAmps with that little dweeb known as _toshiba_ (with the small 't') |
|
|
| tortello |
| quote: | Originally posted by grataku
stefanuzzo,
inerzia si scrive con una -n-. |
Dipende, se fosse elevata... ;) |
|
|
| Fred Dieckmann |
Text in Quotes from moamps post:
"Vunreg drops below Voff=12V, the relays have already reached Vdrop-out value and deenergized."
12 V - 3 V (Vds of Q26 turned on for conducting relay coil current) is 9 V
"I'm telling you that IN PRACTICE Vdrop-out is about 7 V for a 24 V relay"
9 V is greater than the 7 V at which you say the relays actually drop out. So Q25 has muted the relays before the dropout voltage of 7 V you give.
"Switching off relays via MOSFET Q25 is therefore useless because they are already switched off, i.e., Q25 can be left out."
If the relay voltage is 9 V when Q25 mutes, which is greater than 7 V at which you observe that relays drop out, Q25 is opening the relay before the relay coils drop out by themselves. If one designs for the design sheet 2.4 minimum for the particular relay Passlabs is using, one could conceivably have the relay not mute until the supply voltage of 5.4 volts. One designs for actual parameter variation of a large number of parts rather than for a few parts under test in the lab.
You have not specified what relay you are testing with, so one can hardly jump to the assumption about the particular drop out voltage of the relay Passlabs is using. Ironically you gave best reason for including Q25 relay that you said is unnecessary, the adjustability of the power supply voltage point at which the relay mutes via selection of the voltage divider resistor values.
"Now I don't know why I even bothered as you can obviously think of no other argument besides Mr. Pass-knows-best mantra (with all due respect to Mr. Pass) that you keep repeating post after post."
"WhoDaMan !
Thank you Mr. Pass"
I believe I started out with a disscussion of the circuit before expressing the opinion that the circuit worked fine as drawn and had no extra parts, before mentioning Mr. Pass. Hardly a mantra in my eyes. The circuit works and is quite adjustable, result of the design, and not of Mr. Pass's reputation (which is a result of his skill as a designer.) I didn't even know if he designed it at the begging of this thread.
Are you still asserting Q25 is unnecessary? |
|
|
| Jocko Homo |
Someone started the thread saying.....well, we all know what he said.
He is the one confused. I admit, I do not know, or really care who he is. I think we should force someone to apologize for being a twit and inducing yet another rock throwing contest. As entertaining as it might be...............
Everyone else can chose for their own if they feel a need to apologize.
Jocko |
|
|
| moamps |
| quote: | Originally posted by Fred Dieckmann
Text in Quotes from moamps post:
"Vunreg drops below Voff=12V, the relays have already reached Vdrop-out value and deenergized."
12 V - 3 V (Vds of Q26 turned on for conducting relay coil current) is 9 V
"I'm telling you that IN PRACTICE Vdrop-out is about 7 V for a 24 V relay"
9 V is greater than the 7 V at which you say the relays actually drop out. So Q25 has muted the relays before the dropout voltage of 7 V you give.
"Switching off relays via MOSFET Q25 is therefore useless because they are already switched off, i.e., Q25 can be left out."
If the relay voltage is 9 V when Q25 mutes, which is greater than 7 V at which you observe that relays drop out, Q25 is opening the relay before the relay coils drop out by themselves. If one designs for the design sheet 2.4 minimum for the particular relay Passlabs is using, one could conceivably have the relay not mute until the supply voltage of 5.4 volts. One designs for actual parameter variation of a large number of parts rather than for a few parts under test in the lab.
You have not specified what relay you are testing with, so one can hardly jump to the assumption about the particular drop out voltage of the relay Passlabs is using. Ironically you gave best reason for including Q25 relay that you said is unnecessary, the adjustability of the power supply voltage point at which the relay mutes via selection of the voltage divider resistor values.
"Now I don't know why I even bothered as you can obviously think of no other argument besides Mr. Pass-knows-best mantra (with all due respect to Mr. Pass) that you keep repeating post after post."
"WhoDaMan !
Thank you Mr. Pass"
I believe I started out with a disscussion of the circuit before expressing the opinion that the circuit worked fine as drawn and had no extra parts, before mentioning Mr. Pass. Hardly a mantra in my eyes. The circuit works and is quite adjustable, result of the design, and not of Mr. Pass's reputation (which is a result of his skill as a designer.) I didn't even know if he designed it at the begging of this thread.
Are you still asserting Q25 is unnecessary? |
In my post #44 I based all my conclusions on the assumption that the relays were connected in series because, in my circuit-designing experience, I would not normally connect 24V relays on power supply that exceeds 24V. If you take a closer look at the picture in that post, you'll actually see two yellow reed relays connected in series. My opinion is that connecting them in parallel would have caused unnecessary power dissipation, higher magnetic flux, etc., especially in situations when those relays would have to switch signals with very low energy. Besides, in this particular case we're discussing, the relays are connected to unregulated power supply (with possibly significant Vripple), which may induce noise to interfere with switch contacts and for which reason one would certainly want to keep the current thru relays as low as possible. I thought you were familiar with this problem so I had hard time understanding why you found it so silly to implement a pre-resistor in a circuit which was designed in a parallel when that very pre-resistor did the job nicely (i.e. keeping current thru relays at a required level).
You apparently think the relays here are connected in parallel and that's why your calculations are different from mine, naturally. However, Q25 should switch off the relays when the voltage Vunreg drops to about 30V, which doesn't happen regardless of whether the relays are connected in series or in parallel. I have already described how to adjust this thing correctly in my post #47 (long before Mr Pass came in the picture). To sum things up, I don't think this transistor is unnecessary, only that it is not optimally adjusted in the original design. (Please, read my post#50. I hate repeating myself.)
The title of this thread might be inappropriate as it has little to do with what I'm talking about and has caused a lot of misunderstanding as to the actual point of my posts (BTW, thanks Grataku!). As far as I'm concerned, this is my final post on this subject. For those who feel like discussing it further, I'd appreciate it if they started a new thread. |
|
|
| TheMyxin |
[QUOTE]Originally posted by moamps
[B]
Hi,
R77 is 750K resistor, current thru them can be in worst case I=V/R=40V/750000=ca 50uA. Power dissipation on R77 is then P=V*I= 2mW !!!!! , 0,25W is more than enough.
Yes you right. Ok now its a 0,6W Resistor and it works also :-))
Preresistor R19 can be between zero (maximum heating of relays, Q25, Q26) and about 1k (from my experiments).
From my practical experience with 1,5Kohm or 2,2 Kohm for R19 you can recognize heatings on the relais and Q25,Q26. Now with a 3,3 KOhm Resistor for R19 it works also. After 40 Min on ( i have made some other measurements) Q25,Q26 and the Relais are still at room temperature and i cant observe any heatings at this
parts. I have swichted several times on/of and the switching works without any problems. I think from the past that Q25 was really damaged and works sometimes or not and this was the reason for the switching problems. Perhaps due that i have used the Ono without PQuadrats correction on the PCB-Supply-Board.
Many thanks on all posters for understandings for me of this
circuit-part. I have learned a lot from that. |
|
|
| Fred Dieckmann |
On reading schematics:
The Passlabs schematic does not show two coils in series and states that one channel is shown. It was a pretty sure bet that they were parallel.
On reading data sheets:
If you are going to analyze a circuit, READ THE DATA SHEET.
I went through the effort to post a link so people would see how the relay works, i.e., switching voltages, coil resistance, allowable voltages.
Reliability:
The data sheet says that 48 volts (at 50 degrees C) is the maximum voltage, well above the 37 volts the relay actually sees.
Coils in series or parallel?
The data sheet give a +/- 10% spec for the coil resistance. Worst case gives 2592 for one limit and 3168 for the other limit. These coils in series give 16.65 volts of the 37 volts across the coil with lower coil resistance. The spec state at worst case you need 16.8 to turn the coil on. Under these maximum spreads, one of the relays in series will not mute. You have a day where the AC voltage sags (summer in Texas with everyone's air conditioner on) your DC rail will sag a bit and you have even less voltage and be further below the max required 16.8 volts. This is called designing for the worst case tolerance spread. You will might get small enough variations that in the large majority of cases, things work. But sure as hell the tolerances can stack up such that some won't work. Allowing for the tolerance variations and staying as far away from operating area where the cards can stacked against you, is a real good idea.
"However, Q25 should switch off the relays when the voltage Vunreg drops to about 30 V, which doesn't happen regardless of whether the relays are connected in series or in parallel."
I don't know where this number came from. The regulated supplies will track the unregulated minus about 3 volts until the zeners turn on. I will agree that a higher value rail voltage for the mute threshold might be good idea but 30 V is an arbitrary number and the regulated supplies are even a few volts from going into actuall regulation at that point.
"Switching off relays via MOSFET Q25 is therefore useless because they are already switched off, i.e., Q25 can be left out."
Later:
"To sum things up, I don't think this transistor is unnecessary, only that it is not optimally adjusted in the original design."
I guess this means you changed your mind? Actually the thread wasn't a waste: since it went from" it doesn't work", to "it has extra parts", to how it actually works in quite some detail; including how to adjust the threshold, the relay choice and configuration, and how to adapt it to other voltages. I suppose Mr. Pass can forgive us in the end; after a bit of confusion in the beginning and middle. It's a great circuit and should be quite useful (if all the parts are included.....) It shows greater subtlety for a simple circuit than apparent at first glance. |
|
|
| vpharris |
I think you guys need a new circuit to figure out or something. As the youths would say....."Like really, man!"
:rolleyes: |
|
|
| Fred Dieckmann |
| I would settle for people understanding this one. I hate too think of the hand wringing a really complicated one would cause, like maybe a circuit with three transistors ................ |
|
|
|
|
