(AK) classA HA/Preamplifier.

[Deleted]

Jan 7, 2011 12:54:10 GMT @solderdude said:

Jan 7, 2011 8:45:40 GMT @sandyk said:

My preferred option would be to use the attached which is more versatile, and could be used to switch to either headphones or RCA Output sockets IF the constructor felt the need to use such a device.
This circuit is also designed to operate when > +-600mV is detected at the output of the amplifier, as against >1,000mV of the online module.
As Chong will be keeping his voltage regulators inside an external PSU case, he will not be able to use this unit .
It is available in kit form from Altronics.
www.altronics.com.au/index.asp?area=item&id=K5124

This kit does NOT disconnect at +/-600mV but more in the neighborhood of +/- 1.5V which is higher then the very small one Will referred too.

For headphones small signal relays are usually better then bigger relays.
The bigger relays are more suited for LS.

The small PCB can be modified to have fast off too (even without an extra AC signal)

Frans
At least the SC kit is designed to work from the +-20V supply rails.
I gather that you have the schematic of that small PCB ?
I was looking at the SC layout diagram with a view to seeing if a more appropiate DIL DPDT relay could be fitted ,as in my own delay circuit.\All in good time, as the PCBs haven't even been delivered yet !!
As you should have by now gathered , I am not in favour of using a delay/protection circuit with this HA,when used solely as an HA, and do not believe it is even necessary.
There are normally no on/off plops when the JLHs are fitted, in conjunction with the VREGs in the offset correctors.
Too many individual alterations to the basic design by those who don't really know what they are doing is only likely to lead to problems, or at least reduced performance.
I would prefer to see this HA constructed initially in it's original form, as Will and Leo have already successfully done, and then changes made AFTER the basic HA is working correctly.
If people want to add LIghtspeed attenuators or motorised volume controls,delay circuits etc., then they should do that after ascertaining that the HA is performing correctly first.
Alex

[Deleted]

Jan 7, 2011 13:23:17 GMT oohms said:

Surely it would be possible to build an offset detector with a BJT differential amplifier, and adjust the gain to make the difference 0.7v at much smaller voltages?

It would be very easy to design something that can even flip the relay with +/- a few mV but this could result in the amp switching off at certain deep bass notes.

The only things these protect are expensive headphones in case the amp goes faulty.
In many cases this means either the supply rail voltage is on the output or a few volts in case one of the two power supply rails fail.

As Chong has an external power supply it is HIGHLY recommended to use a DC protector.
Also needed when fused DC lines or an JLH with current limiter is used.
remember most of these amps are not put together by electronics experts in which case a fault appearing could be higher then if done by a professional.
I might add, although I am a professional some parts occasionally blow or I make bad solderjoints

The bigger PCB Alex prescribed can easily be made so it does flip at around +/- 0.7V by simply replacing 2 of the 22k resistors (the ones to ground) for 220k resistors.

The small PCB has the same DC protection as that is used in the Panda....
Both PCB's can be used with supply rails between 15V and 60V (with some parts adjusted)

[Will]

On the subject of power connectors (from external traffo to amp), I've read that Deltron silver contact XLR's are used by a man who knows his power supplies from his elbow ;O)

And if you are feeling really flush, check out the Jaeger brand of connectors.

That is all. Carry on!

[XTRProf]

Jan 7, 2011 11:46:47 GMT Will said:

If anybody wants some of these 0.1% resistors for the emitters of the LS devices, as mentioned above, please contact me, as I'm placing an order at the end of the month.

Err, what value without looking at the schematic? I might already have that position 0.1% resistors. Thanks.

[XTRProf]

Jan 7, 2011 13:27:23 GMT FritzS said:

Jan 7, 2011 12:54:10 GMT @solderdude said:

This kit does NOT disconnect at +/-600mV but more in the neighborhood of +/- 1.5V which is higher then the very small one Will referred too.

For headphones small signal relays are usually better then bigger relays.
The bigger relays are more suited for LS.

The small PCB can be modified to have fast off too (even without an extra AC signal)

AMB use one of this
Omron G5V-2-H1-DC12
Omron G6A-234P-ST15-US-DC12
Omron G6A-274P-ST15-US-DC12
Fujitsu RY-12W-K
Fujitsu RA-12W-K
Tyco C93402

www.amb.org/audio/epsilon12/

Wah........, but great! I will try that as well togather with the Panda anti dc offset / delay PCBA. But will that fast turn off as well and using DC? I prefer DC than using AC for obvious reason. Anyway, have you build one yet? If have, any comments?

Thanks Fritz.

[XTRProf]

Jan 7, 2011 13:30:56 GMT @sandyk said:

If people want to add LIghtspeed attenuators or motorised volume controls,delay circuits etc., then they should do that after ascertaining that the HA is performing correctly first.
Alex

No worry, whatever additional will be just in the pipeline. We can discuss future ideas now for the fun, right? Obviously, we will get the AHA to work first. If not working, how get a base to test the additional?

[Will]

Just remember page 3, that's where the circuit diagrams live (it used to be page 4, but something quantum happened)

The values are 4 x 100R and 4 x 150R for the emitter resistors.

[XTRProf]

Jan 7, 2011 13:38:18 GMT @solderdude said:

Jan 7, 2011 13:23:17 GMT oohms said:

Surely it would be possible to build an offset detector with a BJT differential amplifier, and adjust the gain to make the difference 0.7v at much smaller voltages?

It would be very easy to design something that can even flip the relay with +/- a few mV but this could result in the amp switching off at certain deep bass notes.

The only things these protect are expensive headphones in case the amp goes faulty.
In many cases this means either the supply rail voltage is on the output or a few volts in case one of the two power supply rails fail.

As Chong has an external power supply it is HIGHLY recommended to use a DC protector.
Also needed when fused DC lines or an JLH with current limiter is used.
remember most of these amps are not put together by electronics experts in which case a fault appearing could be higher then if done by a professional.
I might add, although I am a professional some parts occasionally blow or I make bad solderjoints

The bigger PCB Alex prescribed can easily be made so it does flip at around +/- 0.7V by simply replacing 2 of the 22k resistors (the ones to ground) for 220k resistors.

The small PCB has the same DC protection as that is used in the Panda....
Both PCB's can be used with supply rails between 15V and 60V (with some parts adjusted)

Hi Frans,

Yes, I agreed that I should be looking at something DC instead of AC implementation for obvious reasons. This will also keep the AC at bay iiterally from the audio PCBA. But is there anyway the Panda dc offset circuit be modified to kick in something lower like 70mv. Or you feel that the modified 0.7V threshold is already good enough for expensive HP protection?

[XTRProf]

Jan 7, 2011 17:17:17 GMT Will said:

On the subject of power connectors (from external traffo to amp), I've read that Deltron silver contact XLR's are used by a man who knows his power supplies from his elbow ;O)

And if you are feeling really flush, check out the Jaeger brand of connectors.

That is all. Carry on!

Heh, heh, heh, nothing from me on this as I firmly believed DIN overall is better as expressed earlier for any power line implemantation. Anyway, powerline don't need super nut like silver contact. OCC copper is already nut enough and super already. Save the silver for the audio side and not the PS side. Just small a change will be detected for PS side.

[Will]

Think about contact resistance, and the effect it'll have on the supply, which is especially important if feeding DC from an external supply.

Can you equal or better 0.0007R with the din plugs?

[Deleted]

The Panda circuit cannot be modified to 70mV.
Personally I think 70mV is too sensitive.
It ONLY needs to kick in on emergencies where the HP's start running a risk of being damaged by a FAULTY amp.
the +1.24V and - 1.04V is good enough.
The +/- 1.5V of the bigger PCB will do too and can even be lowered to +/-0.7V

IF the amp, PS, PS cable or JLH goes faulty the error voltage will be higher then the 1.24V and the protection will do it's job.
The DC servo in the amp itself will ensure there will be NO DC under operating conditions.

Both DC protectors will work fine.
They are virtually the same circuit only 1 is built with discrete parts and the other one with a specialised IC.

If you want it to switch-off fast too (without an additional AC input) contact me when you know the power supply voltages and have the protector at hand you will need to replace a resistor by a (to be defined) zener diode.
You will need to buy a (HIGH SENSITIVITY) 5V relay instead of the enclosed 12V relay with the same properties/pinning.

If all goes well the DC protection will never kick in and only the slow-on - fast off action is used.

[Deleted]

RE DC Protection.
Seems like I am in a "damned if I do, and damned if I don't"
situation here. But hey, what would I know ?
I only had >44 years of working with relays of various types, including mercury wetted Clare relays,reed relays, and many other types of relay. I have seen MANY relay contacts go high resistance due to a lack of a suitable wetting current.
Chong's original request was for no contacts in the signal circuit, and if there had to be relay contacts , they should have "gold" contacts. Friedrich's repeated demands re dramatically increased sensitivity are unrealistic, and he keeps ignoring explanations in this respect from both Frans and myself.That area of discussion has been dragging on since page2 of this thread !
The DC protector used in the Panda DOES have an AC input for fast shut down. As far as I can ascertain, the SINGLE contacts in this delay/protection circuit appear to be nothing too special.
At least in my simplified delay circuit I used 2 DIP8 DPDT relays (one for each channel) with paralleled contacts to help overcome possible problems in this area.
I will purchase an Altronics SC kit and see if it can readily accomodate 2 x 5V DIP8 DPDT relays in series, with paralleled contacts for greater reliability. I would be concerned about long term reliability in that cheap Asian sourced PCB.Neither is it readily useable to switch the output to RCA sockets if required.It also relies on the headphone socket for mounting.
Alex

[Deleted]

Jan 7, 2011 21:30:22 GMT Will said:

Think about contact resistance, and the effect it'll have on the supply, which is especially important if feeding DC from an external supply.

Can you equal or better 0.0007R with the din plugs?

I agree with Will. The contact area of the XLR pins is far larger.
I use quality XLR plugs and sockets for that reason. I don't have any other equipment using XLR cables, so that is not a problem.In any case , cables can be labelled .
Alex

[XTRProf]

Jan 7, 2011 21:18:21 GMT Will said:

Just remember page 3, that's where the circuit diagrams live (it used to be page 4, but something quantum happened)

The values are 4 x 100R and 4 x 150R for the emitter resistors.

Hi Will,

I don't have those 0.1% values yet. So I will take matched to the dot 100R 8 pcs and 150R 6 pieces.

Thanks.

[XTRProf]

Jan 7, 2011 21:34:46 GMT @solderdude said:

The Panda circuit cannot be modified to 70mV.
Personally I think 70mV is too sensitive.
It ONLY needs to kick in on emergencies where the HP's start running a risk of being damaged by a FAULTY amp.
the +1.24V and - 1.04V is good enough.
The +/- 1.5V of the bigger PCB will do too and can even be lowered to +/-0.7V

IF the amp, PS, PS cable or JLH goes faulty the error voltage will be higher then the 1.24V and the protection will do it's job.
The DC servo in the amp itself will ensure there will be NO DC under operating conditions.

Both DC protectors will work fine.
They are virtually the same circuit only 1 is built with discrete parts and the other one with a specialised IC.

If you want it to switch-off fast too (without an additional AC input) contact me when you know the power supply voltages and have the protector at hand you will need to replace a resistor by a (to be defined) zener diode.
You will need to buy a (HIGH SENSITIVITY) 5V relay instead of the enclosed 12V relay with the same properties/pinning.

If all goes well the DC protection will never kick in and only the slow-on - fast off action is used.

Thanks Frans, I will order a few pieces of the Panda dc / delay protection units and get it modified for 0.7V kickin and fast turn off. I will also order the Fritz's link for trial as well to see what it is all about. No try no learn .........

I will pm you how to modify the unit. Thanks.

[Deleted]

Jan 7, 2011 21:18:21 GMT Will said:

Just remember page 3, that's where the circuit diagrams live (it used to be page 4, but something quantum happened)

The values are 4 x 100R and 4 x 150R for the emitter resistors.

Will
I deleted some posts that had no relevance to the thread.
Alex

[Deleted]

Jan 8, 2011 6:30:42 GMT XTRProf said:

Thanks Frans, I will order a few pieces of the Panda dc / delay protection units and get it modified for 0.7V kickin and fast turn off. I will also order the Fritz's link for trial as well to see what it is all about. No try no learn .........

I will pm you how to modify the unit. Thanks.

Only Alex's bigger PCB suggestion (big relay and PCB with discrete parts) can be made to respond to around +/- 0.7V.
The Panda alike (the small PCB with HP out socket) is fixed at +1.24V and - 1.04V.
It could be lowered a bit but not by much.
involves changing resistors (smaller values) and capacitors for bigger ones.
These will not fit on the small board anymore...

I too have seen many relay contacts fail.
Mostly output relays in LS amps and power relays.
Speaker relays cut out because there are too little sparks/current when the contacts are made (which is needed to make proper contact for those kind of relays) and power relays because there is too much sparking (lack of snubbers mostly) when opening burning of the surface relay contacts.
Sporadically (cheap Chinese ?) small signal relays get faulty contacts.
Percentage wise small signal relays are quite reliable IMO.
There are very reliable small signal relays which are hermetically sealed and have a long lifespan.

Theoretically no added relay contacts is better.
But if your expensive HP burns out you wish you had built it in there anyway.
(Alex's damned if I do / don't)

The protection in the Panda DOES have an AC detector but the small PCB with the HP socket on it (uses the same IC as the Panda) appears not to have this feature/connections and requires either modification to accomodate such input or other measures.
As Chong likes to have a separate PS and keep the AC there (prefferred) the small PCB can be modified to switch off reasonably fast with another relay and a resistor changed for a zener diode.
Not as fast as with a separate AC input but fast enough to prevent a 'plop´

[Deleted]

Jan 8, 2011 7:34:46 GMT @solderdude said:

Jan 8, 2011 6:30:42 GMT XTRProf said:

Thanks Frans, I will order a few pieces of the Panda dc / delay protection units and get it modified for 0.7V kickin and fast turn off. I will also order the Fritz's link for trial as well to see what it is all about. No try no learn .........

I will pm you how to modify the unit. Thanks.

Only Alex's bigger PCB suggestion can be made to respond to around +/- 0.7V.
The Panda alike (the small PCB with HP out socket) is fixed at +1.24V and - 1.04V.
It could be lowered a bit but not by much.
involves changing resistors (smaller values) and capacitors for bigger ones.
These will not fit on the small board anymore...

I too have seen many relay contacts fail.
Mostly output relays in amps and power relays.
Speaker relays cut out because there are too little sparks (which is needed to make proper contact for those kind of relays) and power relays because there is too much sparking (lack of snubbers mostly)
Sporadically (cheap Chinese ?) small signal relays get faulty contacts.
Percentage wise small signal relays are quite reliable IMO

Theoretically no added relay contacts is better.
But if your expensive HP burns out you wish you had built it in there anyway.

Frans
I have already suggested the use of "sacrificial" .25W MF resistors in the output area when expensive headphones of low impedance are used.
IMO, anybody who drives headphones to higher power levels than that deserves to have a resistor or 2 go o/c !
Better that, than their ears .
The only problem with that is the possibility of a small charred area of PCB if the right type of resistors are not used.
Do you know of any .25W or .5W MF resistors that are not flammable ?
Alex

[Deleted]

Like these ?
au.element14.com/vishay-bc-components/nfr2500002208ja500/resistor-fusible-2r2/dp/9473971 (2.2 Ohm 0.33W)
or:
au.element14.com/vishay-bc-components/nfr2500001009ja500/resistor-fusible-10r/dp/9473912 (10 Ohm 0.33W)
there are more values that could be used as output resistors (not in the emitter path of endstage)

I have used these kind of resistors a lot as protection in power lines (not in audio gear) and they (depends on the kind of overload) have a small short flame/puff smoke.
These kind is also used in a lot of (Technics/Panasonic) and other brands audio stuff and also I came across a lot of 'spontanious' open resistors without a clear cause.
I replaced them with normal MF but mounted elevated and this works better...
When you mount these resistors elevated there won't be charred area's on the PCB.
Most modern resistors are not that flamable anymore I thought, check the datasheets of the parts if unsure.

Problem with this solution is that when a high-ohmic HP is connected the resistors won't burn and there could be a DC voltage higher then 5V on the output without you knowing it.

Personally I would install a DC protector (Did it in my TX amp, see schematic www.mediafire.com/?c2zqvacgq079dpw ) but I am not as anal as most about parts...
It has a very simple but effective slow-on, fast off DC protector in it with minimal amount of parts and doesn't need an AC input.
Switches off at around +/- 1.2 V but this can be made smaller (+/- 0.7V)

If people are interested in this DC protector (no kit available ) I will post a schematic and tutorial about it.

[Deleted]

Many modern MF resistors burn like a lighted match when their ratings are greatly exceeded. (the lacquer? )
I had a series pass transistor in a Silicon Chip PSU go s/c for no apparent reason recently in the outboard PSU for my 15W Class A amplifier, despite the load being reduced by a factor of 2 due to duplication.
The result was the emitter resistors on one side going o/c after catching fire and charring a small section of the PCB.
After the PSU series pass transistor was replaced , and the amplifier PCB was repaired and resistors replaced, the channel worked normally again. Around 20 years ago I saw similar happen in a different amplifier. That time I saw the resistors burn like a lighted match.
Telstra used flame proof resistors that went o/c. but they were much larger physically.
Alex

[Deleted]

How they go o/c depends a lot on make and the type of overload/wattage.

That's why I prefer a DC protector.. I am also not afraid of the added relay contact resistance in the rather high-Ohmic output of a HP amp when using small signal relays.
But that's arguably subject to subjectivity also.

[Deleted]

Problem with this solution is that when a high-ohmic HP is connected the resistors won't burn and there could be a DC voltage higher then 5V on the output without you knowing it.

Frans
That is possible, but there would be loud clicks on switch on, and the channel would most likely not work.
Neither myself or any of my friends have had a catastropic failure with a preamp/HA in all the many years we have been constructing them. There are no beginners constructing the Class A HA either. Chong is already deciding to do his own thing, so he is virtually on his own if he has performance issues.
Perhaps Chong's own ideas on the subject should be in a separate thread, instead of digressing from the present construction thread ?
Alex

P.S.
Some people even insist on using upmarket silver solder to ensure very low resistance joints.( I don't.) Surely that is no different than wishing to remove any non essential relay or switching contacts ?
You do not believe in subjectivity, but the vast majority of RG members do ! I also believe in using the "K.I.S.S." principle wherever possible.

Personally I install a DC protector (also standard in the Panda) which I think is a lovely add-on WHEN you are already installing a slow on/fast off relay anyway.
It brings only benefits and has NO downsides in this case.

I agree in the case of the Panda, where many VERY inexperienced people are constructing it. Many of them would be doing so without the substantial backup available to RG members from yourself and Mike.

[Deleted]

That may be so...

I think it's for everyone to make up their own minds about it and 'nough is said about this subject (DC protection/ on off relay).

Personally I install a DC protector (also standard in the Panda) which I think is a lovely add-on WHEN you are already installing a slow on/fast off relay anyway.
It brings only benefits and has NO downsides in this case.

Chong likes to build his gear as his 'database' compels him to, but it's his amp/decision/religion which may or may not share views/opinions/religion we have.

Let's bring the subject to the SC AHA again...

[Will]

Hi Will,I don't have those 0.1% values yet. So I will take matched to the dot 100R 8 pcs and 150R 6 pieces.Thanks.

Hi Chong,

I won't be matching resistors, but will just buy some in, and send them onto whomever wants some, this being DIY, not DISE (do-it-somebody-else )

Let me know if you still want the numbers above, or if you want more for your matching.

[Will]

I think it's for everyone to make up their own minds about it and 'nough is said about this subject (DC protection/ on off relay).

Couldn't agree more!

I wish I'd hadn't mentioned that bloody dc board again.

If people want to use them, then that is absolutely fine, but think about it, how many times has anybody on this site had headphones killed by a large dc offset? Compared to the hours of use by different commercial and DIY amps, by all the people who use them, the percentage must be minuscule.