zeno said:
In a way you're right. That is I raised a similar question because I want to use a pcb and still keep connections short.
Apparently a way to do it is to put small caps (around 47uF or so) closer or on the pins themselves, and then larger caps a bit longer.
Some claims have been made on the gainclone being a bit bass shy, and maybe this is related to the cap size. Perhaps starting with 1000uF and then start adding more similar caps will get more bass.
The question is to be alert to where the breaking point with audio quality is. As it has been stated, the original Gaincard used more capacity in the optional Humpty supply.
Carlos
carlmart:
There was a little discussion on bridging a few weeks back, in one of the threads relating to TDA 7294/93 between Robert, Hugh and myself.
It is felt the best way to bridge is to use opamp to invert the signal before the inputs, and run the power amps at normal stable gain in preferred config (inverting/non-inverting).
A complete explanation is provided by Hugh on that thread.
If you can't find it shoot me a mail and I'll dig it up for ya.
Of course the regular bridging is supposed to sound awful (resistor to -ve input of amp2) but I am running a couple of STKs in this way with no apparent issues.
There was a little discussion on bridging a few weeks back, in one of the threads relating to TDA 7294/93 between Robert, Hugh and myself.
It is felt the best way to bridge is to use opamp to invert the signal before the inputs, and run the power amps at normal stable gain in preferred config (inverting/non-inverting).
A complete explanation is provided by Hugh on that thread.
If you can't find it shoot me a mail and I'll dig it up for ya.
Of course the regular bridging is supposed to sound awful (resistor to -ve input of amp2) but I am running a couple of STKs in this way with no apparent issues.
I spent last night adding a 2200uF cap to each supply leg of my prototype. I already had two 1000uF per supply leg, 4200uF total with a 1uF poly cap across the legs. I couldn't detect any increase in bass, but my older B&Ws are somewhat bass shy so some more bass might be there but I can't hear it. When I added the extra 1000uF per leg, I was at my fathers who has some large ESS speakers, the GC has great low end extention but doesn't have quite the "slam" that a couple of our discrete amps has. Another member pointed out earlier in the thread (and I agree) that my GC(with input pots) is not getting any of the benefit from the an active preamp and is probably why the GC lacks that extra low end power.
PS,
Carlos, one of the "other amps" was my DC100 which sounds great. It had the most low end power of the amps we tested but definitely needs some break-in as the highs were somewhat bright and edgy in comparison. My father is going to run it hard and let me know how it settles out. I'll go and get it back next month(if he lets me have it)
.
PS,
Carlos, one of the "other amps" was my DC100 which sounds great. It had the most low end power of the amps we tested but definitely needs some break-in as the highs were somewhat bright and edgy in comparison. My father is going to run it hard and let me know how it settles out. I'll go and get it back next month(if he lets me have it)
.bridging
hi,
when bridging, why don't you guys try the
http://focus.ti.com/docs/prod/folders/print/drv134.html
as input device?????
hi,
when bridging, why don't you guys try the
http://focus.ti.com/docs/prod/folders/print/drv134.html
as input device?????
Output offset voltage.
Carlos,
I see that you have recalculated for Av=22, however you negelected to consider recalculating Rb for proper compensation. There exists the possibility for DC to be present on the output even with the input capacitors because the op-amp's input stage biasing current flowing through external resistors produces a DC input voltage causing DC to appear on the output. I will admit that this is probably going to be a small value (mVs) and with the bridge configuration the effect will be negligible to the load, but I think the key thing here is to minimize the DC on the output so that only the "music" signal remains. Feedback also effects the output offset voltage.
Build it, measure the DC values of the circuit and adjust the components as necessary. I would even consider potentiometers for experimentation until I found values that worked best for your particular application. This would make it easy to balance the gains and to eliminate any "unwanted" DC from the output.
Later,
Carlos,
I see that you have recalculated for Av=22, however you negelected to consider recalculating Rb for proper compensation. There exists the possibility for DC to be present on the output even with the input capacitors because the op-amp's input stage biasing current flowing through external resistors produces a DC input voltage causing DC to appear on the output. I will admit that this is probably going to be a small value (mVs) and with the bridge configuration the effect will be negligible to the load, but I think the key thing here is to minimize the DC on the output so that only the "music" signal remains. Feedback also effects the output offset voltage.
Build it, measure the DC values of the circuit and adjust the components as necessary. I would even consider potentiometers for experimentation until I found values that worked best for your particular application. This would make it easy to balance the gains and to eliminate any "unwanted" DC from the output.
Later,
Re: Output offset voltage.
Well, that's exactly the point I raised on my previous mail. If there might be a way to get that value with some formula.
In the meantime a multiturn trimpot will have to do. Though I don't know if other adjustments will be necessary.
My guess is the inverted side should be DC zeroed first and then get going with the bridge.
Carlos
stadams said:
Well, that's exactly the point I raised on my previous mail. If there might be a way to get that value with some formula.
In the meantime a multiturn trimpot will have to do. Though I don't know if other adjustments will be necessary.
My guess is the inverted side should be DC zeroed first and then get going with the bridge.
Carlos
Sounds like a good place to start...
Carlos,
I think that you have the right idea. By my calculations, the value for Rb should be ~9.5 kOhm. I would go with a 15-turn 20-kOhm potentiometer. And just for fun I would also use potentiometers for both Rf's to see what the effect of balancing the gain precisely would have on the circuit.
Of course this all sounds like overkill, doesn't it? Just point-to-point wire the design and play around with it. That is what I think the beauty of DIY is.
Now, on to power supplies.
Have fun,
Carlos,
I think that you have the right idea. By my calculations, the value for Rb should be ~9.5 kOhm. I would go with a 15-turn 20-kOhm potentiometer. And just for fun I would also use potentiometers for both Rf's to see what the effect of balancing the gain precisely would have on the circuit.
Of course this all sounds like overkill, doesn't it? Just point-to-point wire the design and play around with it. That is what I think the beauty of DIY is.
Now, on to power supplies.
Have fun,
Chipamp PSU versions
For me the question of the psu is not yet solved.
I must state that all the versions I tried are completely hum free (the mere fact how quiet, noise and hum free these amps are is quite amazing). As a matter of fact I tried all version with a single 240VA toroid with two 24V secondaries. I used standard bridges (made by IR) with 8A (4x version) and with 25A (1x and 2x versions).
1. One bridge for both channels
The "Orginal Gaincard" seems to have just one bridge in the Power Humpty only for both channels (according some postings in the Tweakers asylum and the chipamp forum) and no caps whatsoever. But why ?
I tried that with very little positive effect. This was by far the worst sounding version, specially the channel separation was at a very low standard. This fact does not really astonish me, because the possible side effects of the power demand affecting the other channel is obvious (the low valued psu caps of both channels are only seperated by wire).
2. One bridge for each channel
I tried that too (the bridges close to the subsequent channel) and the channel separation is much better and so is the sound too. But both channels must be mounted close to each other because the star ground is then common for both amps (four wires from the transformer, two are going to the common ground, the two others are split to each of the bridges). For that I connected the individual star grounds (where all other individual ground connections are) together with a thick low resistance solid wire.
3. Two bridges for each channel
That is easy to implement, and we have then four individual wires with unfiltered DC to each amp. But then the individual star ground is completely floating and is only connected to the other channel or to the ancillaries via the input ground. That can lead to some higher level of hum (can be solved by grounding then via 10Ohm resistors or paralleld bridges to a common ground or earth).
What is then the main reason for separating the transformer with a significant length of (thin) wire, besides stray effects of the magnetic field ? Or is there a special (filter)effect of the wires (with their resistance) ?
Does anybody hear the difference between fast diodes or standard bridges here?
BTW, bridging the amps add more difficulties here.
Klaus/ - who likes chipamps and enigmas
For me the question of the psu is not yet solved.
I must state that all the versions I tried are completely hum free (the mere fact how quiet, noise and hum free these amps are is quite amazing). As a matter of fact I tried all version with a single 240VA toroid with two 24V secondaries. I used standard bridges (made by IR) with 8A (4x version) and with 25A (1x and 2x versions).
1. One bridge for both channels
The "Orginal Gaincard" seems to have just one bridge in the Power Humpty only for both channels (according some postings in the Tweakers asylum and the chipamp forum) and no caps whatsoever. But why ?
I tried that with very little positive effect. This was by far the worst sounding version, specially the channel separation was at a very low standard. This fact does not really astonish me, because the possible side effects of the power demand affecting the other channel is obvious (the low valued psu caps of both channels are only seperated by wire).
2. One bridge for each channel
I tried that too (the bridges close to the subsequent channel) and the channel separation is much better and so is the sound too. But both channels must be mounted close to each other because the star ground is then common for both amps (four wires from the transformer, two are going to the common ground, the two others are split to each of the bridges). For that I connected the individual star grounds (where all other individual ground connections are) together with a thick low resistance solid wire.
3. Two bridges for each channel
That is easy to implement, and we have then four individual wires with unfiltered DC to each amp. But then the individual star ground is completely floating and is only connected to the other channel or to the ancillaries via the input ground. That can lead to some higher level of hum (can be solved by grounding then via 10Ohm resistors or paralleld bridges to a common ground or earth).
What is then the main reason for separating the transformer with a significant length of (thin) wire, besides stray effects of the magnetic field ? Or is there a special (filter)effect of the wires (with their resistance) ?
Does anybody hear the difference between fast diodes or standard bridges here?
BTW, bridging the amps add more difficulties here.
Klaus/ - who likes chipamps and enigmas
hifi said:
Hi Micke,
Can you explain why "a larger transformer will have higher inductance", please.
I thought that the overall inductance in transformers was more a function of the total number of turns of wire used (for the secondaries) in manufacture, which will be far more influenced by the output voltage than anything else, surely.
I confess to not having carefully researched the construction of transformers, but does a higher VA rating *always* mean using a greater number of turns of wire, or can it be done by using difference gauge wire?
I wonder if, perhaps, you really intended to say "higher voltage" here, rather than "larger", which I understand to mean a higher VA rating?
Many thanks for clarifying this point for me.
Regards,
hello,
The number of turns will remain the same for different VA ratings (for a constant voltage), but the Core will not, the larger VA rating the larger the Core, witch will result in that a longer Copper Wire must be used to wind the transformer,witch will result in a higher inductance.
The gauge of the wire will also affect the impedance but not as much as the inductance, the thicker copper for larger transformer is mainly for greater power dissipation.
/ Micke
The number of turns will remain the same for different VA ratings (for a constant voltage), but the Core will not, the larger VA rating the larger the Core, witch will result in that a longer Copper Wire must be used to wind the transformer,witch will result in a higher inductance.
The gauge of the wire will also affect the impedance but not as much as the inductance, the thicker copper for larger transformer is mainly for greater power dissipation.
/ Micke
inductance
As far as I remember, the larger the core, the *lower* the number of turns per volt, because the core can take more flux.
That aside, I don't think one can consider the xformer as an inductance as far as the load is concerned.
Observe: if the rectifier diodes are conducting, there is almost a short, which is transformed to the primary as, you guessed it, almost a short. That is why there will be very high current pulses, 10-s of amps, I have seem 100A peak on some large reservoir caps. So at that point there is no inductance, because otherwise there would not be such a large current.
A transformer loaded at the secondary only has the stray inductance, which is a result of the less than 100% coupling between primary and secondary.
Jan Didden
As far as I remember, the larger the core, the *lower* the number of turns per volt, because the core can take more flux.
That aside, I don't think one can consider the xformer as an inductance as far as the load is concerned.
Observe: if the rectifier diodes are conducting, there is almost a short, which is transformed to the primary as, you guessed it, almost a short. That is why there will be very high current pulses, 10-s of amps, I have seem 100A peak on some large reservoir caps. So at that point there is no inductance, because otherwise there would not be such a large current.
A transformer loaded at the secondary only has the stray inductance, which is a result of the less than 100% coupling between primary and secondary.
Jan Didden
hifi said:
Hi Micke,
I still don't fully understand what is going on here, I am sorry to say, and I wish I could get my head around precisely what it is you are saying.
I realise that every piece of wire will have some inductance, and that this will be related in a linear fashion to its length.
However, I always though that, compared with the inductance caused by the 'coil' winding (i.e. because it is in effect an inductor) any inherent inductance due to the wire, itself, would be miniscule.
Why I am especially interested in getting to bottom of this is that I have spent many years 'listening' to different components in audio circuits, with a view to using the best from a 'sonic' point of view.
I confess that until your post about inductance in transformers causing a 'slowing' effect, I had never considered comparing transformers from their inductance aspect, so to me, this opens up a whole new 'can of worms'!
Clearly, as I said before, the output voltage will affect the overall inductance (due to the different number of secondary coil turns in the transformer) in a relatively linear fashion, and doubling the output voltage would be likely to have (approximately) twice the number of secondary turns, and (approx.) twice the resultant inductance, I would have thought.
Having seen elsewhere that someone has decided that a 12volt secondary transformer gives the best 'sonic' results in these Gainclones, if inductance does have some (adverse) affect on this issue, would it not be more likely that it is the substantially lesser number of turns in the 12v transformer secondaries which is the main cause of this, rather than any (very slight?) change in inductance due to any decrease/increase in wire length with a lower/higher VA rating?
Of course, the DC impedance of the secondaries will also be affected by the length/thickness of wire used, but this is a rather different matter.
Regards,
Inductance for given primary and secondary voltages is going to vary greatly depending on core topology, size, and material. Although the /ratio/ of secondary to primary inductance will be determined by the secondary to primary ratio, the actual inductance will depend on the transformer you use.
Regardless, I don't understand how the inductance can have anything to do with the "speed" of the supply. As janneman noted, only the primary leakage inductance is seen as inductance to the supply, and this again is more a function of design than of voltages.
Regardless, I don't understand how the inductance can have anything to do with the "speed" of the supply. As janneman noted, only the primary leakage inductance is seen as inductance to the supply, and this again is more a function of design than of voltages.
What happened to Brian and Jordan's experiments and comparisons of their gainclones? Some other people as well don't seem to have "reported back" on their gc's.
My impressions are still extremely positive after a few weeks of listening to my gc.I bypassed the main 1000uf caps with 2,2uf film caps and it seems to me that the sound is a bit more powerful in the bass. I think the soundstaging is excellent and is really ahead of the Aleph 5 .This is particularly obvious with any live recorded music or with classical.I am still amazed how smooth it sounds at high frequencies.There seems to be more detail without sounding etched or bright.
My impressions are still extremely positive after a few weeks of listening to my gc.I bypassed the main 1000uf caps with 2,2uf film caps and it seems to me that the sound is a bit more powerful in the bass. I think the soundstaging is excellent and is really ahead of the Aleph 5 .This is particularly obvious with any live recorded music or with classical.I am still amazed how smooth it sounds at high frequencies.There seems to be more detail without sounding etched or bright.
Alright.. We haven't done extensive testing, but of the two amps we currently have, the one with the Auricaps and rikens sounds a lot better/brighter/more defined than the one with the solens and allen-bradley resistors. The allen-bradley/solen amp was the only one checked on the oscilloscope; I'm pleased to say it was pretty stable (since Brian tested it, you'll have to ask him to quantify "pretty stable").protos said:
I'm extremely satisfied with these amps thus far, but I don't want to say too much about them until I've been able to do more conclusive testing. For all I know, the auricap/riken amp sounds brighter because I covered all my pcb traces with silver solder.. Another concern: I don't know what's going to happen when I replace my regulated power supply with a transformer.
Oh, and speaking of pcbs, I threw together a board for the inverted lm3875 a week or two ago.. I've attached a picture of the design.. I'll post some pictures of the actual board a little later. The goal was to keep the wiring as short as possible. The chip mounts underneath the board, the pins have to be bent up. The caps sit on the top of the board.. The NFB resistor is the only point-to-point connection, underneath everything and resting just above the heatsink, soldered to 3 and 8.. I heard once that thermal tracking of components is a good thing--anyone know anything about this? I ask because I could have made it so the resistor was actually resting on the heatsink. There is a star ground for each channel.. I don't know if that's a good thing or not.
Anyhow, I'm still messing around with these chip amps.. I'm just reserving final judgement until I've done more testing.
Oh, and regarding my radio problems, I placed my chip amp in a cage which use to house a redundant power supply for a server, and it knocked out most of the noise... I'm in the progress of putting together an aluminum case and a steel/copper case right now. Once they're done, I'll compare the shielding.
--Jordan
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.