| PauSim |
This message was posted as a reply by accident so Iīm reposting as a new thread as originaly intended... sorry for any inconvenience.
Greetings to all inmates
Itīs been a month since I started to read this forum with great interest and amusement. I learned a lot here.
Folowing Thortenīs suggestion, I decided to build a GC running on batteries.
I know lm3875īs are specified, but 1875 are far easier to obtain. So...
-Basic NIGC 1875 setup: 47k log pot, 2.2mF cap, 22k res signal-to-ground; 22k + 680 ohm res on the NFL; 1000mF caps on the power rails; no Zobel.
-Two 12v, 7.2AH batteries in series per channel; all grounds merging to the point between batteries (0 volts, I guess).
It should sound great but everytime I turned it on I got this overdrive-type distortion that ruined almost all my speakers even at low volume.
Running on common 9v batteries or even on 4,5v 3r12īs (you know the type?) it does sound great indeed, but they donīt last long enough, of course, so the circuit is well built.
Shouldnīt 1875īs accept 25v per rail (thatīs 25w power) or is 7.2AH too much capacity? I plan to try 3875īs as soon as theyīre available here.
What must be going wrong? Please help!
Thanks
:eek: |
|
|
| GregGC |
First thing to do: Turn it on (with the problem batteries) with input connected to a source but no signal comming out of the source and measure the output DC offset, + rail to GND and - rail to GND. Ones we have that we'll see what you need to do.
Thanks. |
|
|
| PauSim |
Hi GregGC
On the left channel:
39.5 mV
On the right channel:
32.1 mV
these are typical values arenīt they? |
|
|
| PauSim |
Hi, I think Iīve found the cause
Today I fed one channel via a small dual secondary transformer (230V/6+6V 12VA). It sounded ok until I interrupted the 0V connection. I heard the same distortion (although much more pronounced) that happens when the amp is fed by the 12V SLAs.
It seems like connecting the ground star to the point between batteries is not working.
Searching through the solid-state forum, I found that the inclusion of a TLE2426 rail-splitter might solve the problem. Iīll give a try tomorrow and post the results.
Searching through google also found this:
It still puzzles me why Iīm facing this problem. After doing a simple NAD PP-1 conversion I really thought that setting up a gainclone like this wasnīt that complicated. This might explain why noone yet thought about posting a battery Power supply schematic here or in a website (GregGC, NUUK?... )
Hope this helps anyone who may be experiencing the same difficulties.
See you tomorrow :apathic: (cross-fingered) |
|
|
| Morse |
Hi PauSim;
Ummm, I could be totally off-base on this, but do you have a voltage divider consisting of two identical resistors in series (chosen for maybe 1 or 2mA of draw), running from V+ to V- with the centre tied to ground? I wouldn't try tying the midpoint of the batteries directly to ground, since the two batteries may not be matched as closely as you think they are. Also, I would take a pair of 1000uF caps and put them in series from V+ to V-; their midpoint would be tied with the midpoint of the voltage divider you've made, and thus to ground.
That's the way I've always built battery powered headphone amps anyway. (*Knocking on wood* for luck) Thus far it's worked fine here.....
Good luck and all the best,
Morse |
|
|
| Nuuk |
| quote: | | This might explain why noone yet thought about posting a battery Power supply schematic here or in a website (GregGC, NUUK?... ) |
Hi Paulo, the reason that there is no battery driven Gainclone design on Decibel Dungeon is that I have not built one. As a rule, I tend to only include advice based on my own experience.
I would like to try a battery supply myself but that project is in a long queue of other things which require my meagere financial resources! ;)
I have a gut feeling that what is happening with your set up is that one of the batteries is failing causing the rail voltage (or some of it) to appear on the speaker terminals. Have you checked all four batteries? |
|
|
| PauSim |
Thank you so much for your help. God, I love audio electronics but the lack of some basic knowledge drives me crazy.
morse:
You were right. No two batteries werenīt matched. Why didnīt I think about it before? If one battery measures 13.04V and another measures 13.08V of course the middle point wonīt be 0V, but the difference between the two, and in this case in the order of the tenths of Volts! No wonder I damaged so many ("multimedia") speakers.
For my next project I will invent the wheel.:dead:
No, I do not use any kind of voltage divider and Iīll follow your advice soon.
What resistor value do you recommend? Sorry for asking.
Itīs plain Ohmīs Law isnīt it?, my "calculations" for 2mA draw say
12V/0.002mA=6k but seeing the last results I doubt it is so linear.
nuuk:
Please pardon my clumsy English if I looked rude or anything less polite for asking for some schematics of battery power supplys posted on your excelent site.
You see, most people talk about how wonderful batteries sound (and theyīre right), there are schematics of smart battery chargers everywhere on the net, but to see an actual drawing of the setup, "nicles batatoides" as we say here. A "battery power supply for dummies" faq would be a welcome addi(c)tion;).
Too many people say itīs so easy ("simply plug it here and here and voila" etc) but to me itīs just another myth, a fountain of trouble if youīre a novice.
If my batteries were a car engine it would be a Twin Spark.
Last night while working on a circuit I accidently shorted one of the batteries with a piece of 1mm silver wire that got stuck #$%&. It didnīt burn like cat 5 wire. Instead it broke in two pieces, and one of them soldered itself to the plus terminal! And what a damned pretty solder joint it was. All without WBT.
What am I posting here? I should sell the idea to Siltech :devilr:
Well tomorrow will be D-Day. Thanks for the help, guys.
:apathic: (fingers crossed) |
|
|
| Morse |
Hi Paulo;
Sorry to read that you've damaged some speakers during testing...been there, done that myself. One thing I'd suggest you look into are a pair of the cheapest drivers you can find - then use them for testing. I use a pair of $1 each Radio Shack car speakers I got on closeout as my 'testing' speakers. If (or when) they fry, I'll not shed too many tears, though I'll have to find some more el-cheapo speakers....
>>>...No two batteries werenīt matched. Why didnīt I think about it before? If one battery measures 13.04V and another measures 13.08V of course the middle point wonīt be 0V, but the difference between the two...<<<
Actually with batteries it's worse than that, since you're not measuring them under load. Each battery has it's own "internal resistance" which will increase as it ages (when a battery is "used up", it's internal resistance will be maximal). Since each battery has it's own internal resistance, the voltage it produces under load may vary significantly from it's unloaded voltage, which may result in far greater differences than you measured.
For the voltage divider, yes, use Ohm's law. So if you've got 26V from V+ to V- figure that for 2mA you'll have 13000 ohms for your two resistors in series, or 6k5r each. Since 5k6r is a standard size, you might try that. On my last TDA2030A amp, I used a pair of 2k2r resistors in series as bleeders for my PS caps, yielding 8.6 mA across each. Since I used a conventional PS trafo instead of a batttery, I didn't care about the energy wastage (and wanted my PS caps to discharge quicker!).
Definitely put in the PS filter caps. Even with a battery you'll get some ripple due to the changing internal resistance of the battery (at least that's my understanding of the effect). A pair of 1000uF, 35V rated caps would be fine and will be cheap (you don't have to go with any fancy name brands if $$ is tight!).
Good luck! I hope your amp is soon up and running!
All the best,
Morse |
|
|
| Nuuk |
| quote: | | Please pardon my clumsy English if I looked rude or anything less polite for asking for some schematics of battery power supplys posted on your excelent site. |
Hi Paulo, no offence taken and I agree with you, it is frustrating to see something talked about as 'simple' when some of us have no idea how to implement it. I know it takes time but a schematic is the best way to show off any idea.
Just to back up what Morse said. When testing any amp (I tested a new Gainclone NIGC just now), I measure the DC offset and then connect up some old speakers before I connect up any speakers which I want to keep. ;) |
|
|
| dayvel |
| You may even want to consider running the 1875 from a single ended supply. There's a schematic for this in the National app notes. |
|
|
| PauSim |
-Used 2x4.7Kr resistors in series across -25V & +25V.
-Shunted -12V to +12V batteries terminals.
-Tied ground wire to middle point (between resistors, not batteries).
-There are already 2x1000uF/50V caps from power rails to ground, next to pin 3 & 5 (LM1875). Linked 2x100uF caps in series and paralel to the resistors where these are, right next to the batteries terminals, just in case.
It plays for about 2 seconds, after which it silences, just like when you connect only the + and - and no ground. Caps having nowhere to discharge to?
Monday morning Iīll hunt for a TLE2426 and give it a last try:mad:
Guess what? These 80 cent 3R12s are still powering my gainclones, so in the meanwhile Iīll be listening to some glorious music:D
Thanks for your help and goodwill. |
|
|
| PauSim |
Remembered that the voltage divider scheme doesnīt work with 3R12s either. Please check this if itīs correct.
(Sorry for the Picasso style). |
|
|
| Morse |
Hi Paulo;
Sorry to read you're still having problems; unfortunately I do not have a programme to unzip that file so I can't read the schemos.
The caps should be wired so that the (+) lead of the "top" cap goes to V+ and the (-) lead of the "bottom" cap goes to V-. The join of the two caps should be tied to the centre point of the voltage divider, and this should go to ground.
Assuming that you've got things wired this way, next up, have you tried hooking a DMM to the unit and checking the voltages both before and after it functions?
From what you're describing, it sounds like an internal protection circuit in the chip is being activated. It could be that one of the batteries is "misbehaving" so badly that even with a voltage divider you're dropping below the min voltage under load, or some other such thing. Are you using a "deep cycle" battery or an ordinary automotive type battery? Auto batteries are not designed for deep cycle operation, so you'd be better off with something like an RV battery or perhaps a golf cart battery.
Hope this proves useful. Please keep us all posted on how it's going!
Good luck and all the best,
Morse |
|
|
| PauSim |
but there are far more components, which IMHO could compromise the sound, and would also go against the spirit of the gainclone for that reason.
Well, this is supposed to be my "official" opinion, but
to be honest, Iīll consider that, really, since what turns me on is the Final Labs Music-6, NOT the 47 Gaincard.
Iīve read all the available reviews, including the pros and cons, and what counts after all is the sound, and it seems the Final amp wasnīt designed the same way.
I remember one reviewer saying that the ampīs not grounded at all, which makes me suspect that a single supply is employed.
Cheers
:) |
|
|
| PauSim |
The batteries are Yuasa style. Although cheaper, they were purchased through respectable retailers in Lisbon. I got four of them and always switch batteries to check misbehavings.
I use my Wavetek 15XL DMM for measuring and everything seems to be ok.
Iīve read somewhere about mismatching chips so Iīll try switching chips too, if the TLEs wonīt work.
Thanks:) |
|
|
| A 8 |
I have not read all but noting your talk about virtual ground is wrong in an power amp, you need to connect the 0 volt battery terminals to your signal ground for it to work.
It should be very straight forward, you have a total of four batteries that you should connect in series and pick your ground in the middle and the +- 24 volts at both ends.
If this does not work with your powersupply ground and signal grounds connected something else is wrong (provided that they are charged and actually give you +-24 volts). |
|
|
| tiroth |
| quote: | Originally posted by A 8
I have not read all but noting your talk about virtual ground is wrong in an power amp, you need to connect the 0 volt battery terminals to your signal ground for it to work. |
Yes! All this talk about ground tracking the supply midpoint is rubbish. The National chips do not require tracking supplies. It sounds like a grounding issue. |
|
|
| macboy |
| quote: | Hi PauSim;
Ummm, I could be totally off-base on this, but do you have a voltage divider consisting of two identical resistors in series (chosen for maybe 1 or 2mA of draw), running from V+ to V- with the centre tied to ground? I wouldn't try tying the midpoint of the batteries directly to ground, since the two batteries may not be matched as closely as you think they are... |
They do not need to be matched. The ground point absolutely does not need to be at the exact midpoint of the positive and negative supplies. Note that at any given time, the amp only loads one rail anyway, so that one will sag, so the "ground" is virtually never at the exact midpoint, regardless of power supply design (batteries, transormer, whatever).
The important thing is that the signal is referenced to this ground, and that the output is also. The speaker is connected to this ground, and the input should have a cap followed by a resitor to ground (47 kohm is common; anything from 10k up to 100k should be OK). That will DC-reference the input signal to ground.
The because the current through the speaker returns through the ground, the ground must be low-impedance. That means that using a simple voltage divider is not sufficient. It is OK for the headphone amp, especially when used with some caps to decouple the AC current from the resistors, but it is not sufficient for a loudspeaker. |
|
|
| Morse |
Hello Macboy;
>>>...That means that using a simple voltage divider is not sufficient. It is OK for the headphone amp...but it is not sufficient for a loudspeaker...<<<
No offence, but I can't agree with that - I'm listening to an amp right now that uses a voltage divider to establish ground in between V+ and V-. It's driving a set of loudspeakers rather than headphones, and is doing rather well at that. Why can't you use a voltage divider here? I hope the reason is not persuasive enough to make my TDA2030A pack it in! ;)
>>>...especially when used with some caps to decouple the AC current from the resistors...<<<
???
The caps are there to even out any ripple from the battery (since it's voltage out is a function of the loading and changing internal resistance) not to decouple AC from the resistors. In other words, the caps act as a reservoir, just like in a "regular" power supply (though without the 120cps ripple to contend with).
You'll note that I said I was merely guessing as to the reason for his difficulties. Frankly chips aren't my usual fare - but they're wonderful fun to play with and they can sound amazingly good for their low $$ pricetag.
Anyway, since Paulo has gotten the thing working with a different battery, and this one is proving problematic, I suspect that the issue is not in his amp's circuit, but rather in the PS filtering (i.e. the amp's interface to the battery). Still, that's just a guess.
All the best,
Morse |
|
|
| A 8 |
None of us can say why it works unless you give us your schematic but my guess is you got your return current routed through your chassies or some other NON controlled path.
Whatever It is NOT the way to design a good power amp.
think about it....your signal is relative to ground and the current needs to flow back to the batteries and you want that done without additional losses.......stuffing resistors in a poweramp supply to create a virtual ground is really complete ****. |
|
|
| runebivrin |
Actually, the "ground" path for the loudspeaker will be decoupled by the capacitors that are parallel to the voltage divider resistors, so it will kind of work, but my understanding is that capacitor coupled loudspeakers are a historical anomaly.
Rune |
|
|
| Morse |
Hi A 8 and runebivrin;
>>>...None of us can say why it works unless you give us your schematic...<<<
A schemo? Simplicity itself; the PS trafo runs each secondary lead to a bridge (made of 4 discreet, snubbed, FR diodes). The bridge's (-) lead ties to the V- leg and it's (+) lead ties to the V+ leg. With battery supplies, I just leave the batteries float and attach each end of the stack to the appropriate rail. You do the same thing with a trafo with respect to ground - the secondary's centre tap ALWAYS is left floating in this topology.
>>>...but my guess is you got your return current routed through your chassies (sic) or some other NON controlled path...<<<
Not at all. My return from the divider is through a separate 18 ga. ground shackle to the star grounding point. Only thing I use chassis ground for is the metal shield around the potentiometer and the metal chassis of the power switch. All other grounds are discreet.
>>>...current needs to flow back to the batteries and you want that done without additional losses...<<<
You'd better get rid of the LED indicator lamps then. Most people bias them for 10 to 20mA current flow. I only suggested a 2mA current flow for a divider.
>>>...stuffing resistors in a poweramp supply to create a virtual ground is really complete ****...<<<
Every high voltage linear supply out there uses resistors in the PS. They're the bleeder resistors, which are essential for safety. A voltage divider in a PS functions the same as a bleeder resistor, only with a centre tie point. Why not use it to establish ground?
Hi runebivrin;
>>>...capacitor coupled loudspeakers are a historical anomaly...<<<
Actually I am not capacitor coupling my outputs, though it's still a common practice to do so on White Cathode Followers. The confusion arises, I suspect, because there's an implicit assumption that I'm using a single supply. Not at all! This is a split supply, with ground established by the divider and very little DC present on the output (around 2mV on one channel and 3mV on the other).
Ja mata,
Morse |
|
|
| unitgain |
| quote: | Originally posted by Morse
Hi A 8 and runebivrin;
>>>...current needs to flow back to the batteries and you want that done without additional losses...<<<
You'd better get rid of the LED indicator lamps then. Most people bias them for 10 to 20mA current flow. I only suggested a 2mA current flow for a divider. |
Hm. We're talking about signal currents here, don't we?
| quote: | Originally posted by Morse
>>>...stuffing resistors in a poweramp supply to create a virtual ground is really complete ****...<<<
Every high voltage linear supply out there uses resistors in the PS. They're the bleeder resistors, which are essential for safety. A voltage divider in a PS functions the same as a bleeder resistor, only with a centre tie point. Why not use it to establish ground?
Morse |
The current flow doesn't magically end at the ground, it has to go back to the PS. How does your signal current, the one that goes through your speaker, flow back to the PS? Through a resistor? |
|
|
| PauSim |
Guys,
Iīve been following your posts with great attention and tried everything you advised regardless the current debate.
Sadly the problem continues, and before returning it to the devil Iīll you HOW the entire setup is built. Maybe Iīm missing something crucial.
Source Rega Planet 2000 cd player, no DC output.
Currently listening through improvised speaker arranjement: creative labs sub box containing only woofer, on top Monitor Audio monitor two gold with midbass unit (one of the victims, sended to repair) removed.
All cables are home-made: silver ribbons a la goertz for the tweeter (high capacitance), poormanīs nordost 34 conductor flat cables for the woofer; twisted pair silver/teflon interconnects.
So far none failed tied to an ASL SET.
Amp: NIGC; lm1875; basic NatSemi datasheet except 22k/680R; 1M omitted; 47k cheap log pot; sole caps are two 1000/50 are used instead.
Both channels mounted separetely on perforated bread boards, continuous straight tracks. Still a prototype, no box. Everything on-air. Components occupying the least possible space, fb 22k next to IC. Component values matched left/right wised. I forgot
to match fb 22k with signal 22k, difference about 0.2R.
All resistors are common carbon 1/4w except zobel 1,5R- 2watt.
lytics and zobel polyester are Philips.
Only two bridges per channel, from lytic caps to groun. From breadboard run hook-up single-strand cat5 wires to bateries +, - middle-point; signal female rcas and output speaker posts and vol pot. ground star is plain straight track from which the cat5 single strands go. Wires are tied to batteries with croco clips. Originally middle point was wired with single silver wire from +to -and single cat5 strand to ground star. changed that to two cat5 strands shunting + and - only in star ground track attempting to reduce wire resistance. |
|
|
| PauSim |
Always two batteries in series per channel were employed. 12v per rail in reference to ground, not 24 as I first stated, lending to some confusion. Thanks A8.
Gotta go. Nightshift, this week. Sorry
Please do not reply until this post is finished cause thereīs more to come.
Thanks! |
|
|
| Morse |
Hi Paulo;
I know you asked us to hold off replies 'till later (sorry!), but there's one question I wanted to ask - you were VERY detailed in your description, but you didn't mention heat sinks. Did you use them? If not, the higher input voltage of your "problem batteries" could be generating enough heat on the chip die to trigger the automatic shutoff. The "good batteries" may be just low enough in voltage that you're on the "barely acceptable" side of things. Just an idea.
Good luck!!
Hi unitgain;
>>>...Hm. We're talking about signal currents here, don't we?...<<<
Not at all. This is the power supply end of things. If the V+ and V- have a ground established at exactly the halfway mark 'twixt the two, it's my understanding that these wunderchip thingies will use that as the reference for the signal. Otherwise you need to use coupling caps at both the input and the output to avoid excess DC levels leaking out (just like the way most valve amps couple the stages together). In any case, the current through the voltage divider is between V- and V+; since the PS on ss amps is usually good for a couple of amps, who's going to miss a couple mA?
>>>...How does your signal current, the one that goes through your speaker, flow back to the PS? Through a resistor?...<<<
No. Like everything else it has it's own discreet ground shackle to the star ground. Signal in also has it's shackle, though I may pull that and go to a chassis ground there, since the main star is VERY close to the RCA in and I can always use those insulating grommets elsewhere.
All the best,
Morse |
|
|
| mikelm |
sorry, but reading this had been almost painful
please just try it like this - it's simple
the 0.1uF caps need to be close to the IC pins
you may need a cap in series with the input.
good luck
mike
p.s. it does not need to be connected to earth but it may produce a marginal improvement in the sound.
YOU MUST USE 5 -10 AMP FUSES ON THE + & - RAILS RIGHT NEXT TO THE BATTERY TERMINALS. |
|
|
| Morse |
Hi Mikelm;
*Ack!* You beat me to the 'enter' key. ;)
I just noticed the word "sole" in front of caps myself. Yep, if he's missing the .1uF (preferably ON the pins - can't get closer than that) then it's got problems, regardless.
All the best,
Morse |
|
|
| runebivrin |
Ah, how 'bout a little look at Kirchoffs laws...
The current that flows into the loudspeaker comes out at the other end and continues through... what, exactly?
Don't tell me it's the magic of the 3875. It continues right through the decoupling capacitors across the voltage divider resistors. Given large enough capacitors, you'll have a reasonably low roll off frequency in the bass range, but the usual theories on capacitor distortion still apply.
Rune |
|
|
| unitgain |
| Morse, there are a couple of whole Amps flowing through your speaker at full volume. That's a whole bunch of electrons per second! Do you think they all gather at the point you arbitrarily call "ground" and just hang there? |
|
|
| ronc |
Any probems with battery P/S get with me out of forum at roncla at txucom.net
naturally the at=@
ron |
|
|
| Morse |
Okay, runebivrin;
Time (for me) to eat my words (at least they're low carb).
Yes, a voltage divider is indeed in the return path of a split supply so with a really high powered amp, it could be problematical (i.e. small resistors could go 'pop' and have a potential for introducing noise into the PS). Since I regard a 16wrms amp as being almost overkill, it's not been a problem here....but we don't all use 96dB SPL speakers.
However, it does work though I'd suggest a direct return to centre on the trafo (or battery stack) in addition to the resistors. After retrofitting it to the TDA2030A, and testing a. it works and b. there's no change in sound quality that I can here.
I will stand by the necessity of putting bleeders on the PS caps; it's a matter of safety.
Thanks for pointing out the flaws in my reasoning.
All the best,
Morse |
|
|
| mikelm |
I think we can assume that morse is not deluding himself and his amp is working so any amount of theorizing about why it dosn't work seems a bit pointless
However, I think you may find morse, that the bass may improve if you connect the speaker directly to the midpoint of the supply if that is an option with your PSU.
A potential divider and caps could be used at the input to define the midpoint there and if one of the resistors was variable this could be used to trim the dc offset across the speaker to zero.
Pau Sim
If you still have problems it may be that your high capacitance wire is upsetting the chips.
keep us posted..... |
|
|
| unitgain |
Morse,
once again, an Amp or two must return from the speaker to the PS. Do you think it all goes through a 100K resistor? What kind of power would it dissipate? P=I^2*R, you know.
Or let's try it this way. Your resistor is in series with your speaker. The speaker is 8 Ohm. The resistor is 100K Ohm. Don't you think there's something fishy goes on here? |
|
|
| protos |
| I am one of the few with a battery powered GC and have absolutely no problems that you mention.Perhaps if you post a schematic of your connections it will be easier to help you.A tle is not the solution here. |
|
|
| Morse |
Hi Unitgain;
>>>...once again, an Amp or two must return from the speaker to the PS...<<<
Only at full power - my least efficient speakers are 89dB SPL and my most efficient experimentals approach 100dB SPL - I seldom use more than a watt during peaks, and well under 1/4 a watt for average listening. Thus the amount of current is rather less than an amp or two. Let's say that you really are returning a full amp or three though.
Now, obviously the cap is providing a preferential pathway over the resistor, due to Xc = (1/2*pi*freq*capacitance) < resistor's value. Have you ever have a cap fail due to excessive AC ripple? I have. Then we have a dead short here - bad form all 'round in my book.
For the resistor and it's chance of failure, have you ever had a cap come loose from it's connecting wire inside the can? I have (repaired a buddy's speakers a couple years back where that happened inside the crossover - it took a while to figure out, since all the solder joins were good and none of the parts LOOKED damaged...). If that happened as I built it originally, then it wouldn't just be the loss of a reservoir cap - it would force any return current through the resistor. Since the resistance in the line would increase, obviously there would be a corresponding drop in the output power. However, I think it comes down to the size and power rating of resistor used, as to whether or not it's in danger of popping. For low voltage minimum space situations, I use the smallest parts I can get away with - and those resistors can be as small as 1/8 watt if I'm only planning on having them see an mA or two.
Without running the math (and being rather bushed after spending a couple hours tracing the current paths inside a TDA from it's spec sheet, convincing myself that indeed this had to be the return path, then experimenting with this PS) I posted a general caveat against what appears to be a potential problem spot. If the numbers later show that even a 1/8 watt resistor is "safe" here, then that's good news.
Finally, after spending the morning listening to the "new and improved" amp, I can state definitively that there are no audible differences of any kind - at least using any of the speakers I have on hand and using the headphone adaptor with every set of cans (save my Senn 600's). The PS trafo does run marginally cooler though, and that's always welcome.
Did it work as originally built? Yes. Are there audible differences? No. Do I regard the old system as "foolproof"? Well, no. I can see circumstances in which I'd like the return path to be allowed through a CT on the secondary and will use this as my default design from here on out. Thank you again for posing the question as you did Runebivrin; you made me think (always dangerous).
However, for small power amps using non-CT secondaries the original system will work fine. Just make sure that your PS caps can handle the ripple and don't use a midget resistor for the voltage divider. For small watt amps using non-CT secondaries that must use a split supply I will indeed continue as previously.
All the best,
Morse |
|
|
| Morse |
Hi Mikelm;
You could be right about a potential for better bass by shackling the secondary's CT, since the PS caps are in the return path without that.
In my case there are no audible improvements (including bass); however I don't build any speakers for the sub 40cps region any more (it's a matter of practicality - large, efficient bass speakers are loads of fun, but I have a small apartment these days). The main speakers I'm using on this system are a pair of Radio Shack 40-1354's in 12L BR cabs; free air f3 is around 57cps, though I'm wall loading them which definitely helps the low end. Anyway, it could well be that the bass DID improve at the extreme low end - and it's below my speaker's f3.
I definitely do like the idea of a pot for adjusting DC offset here; that's essentially the same sort of thing you'd do with a DHT's filament (the cathode resistor would go to the wiper) to cancel AC hum. DC offset has not been a problem with this amp though. Both before and after shackling the CT on the secondary, it measures 2mV in one channel and 3mV in the other. I attribute that to hand matching the resistors to under 1% before use.
Thanks for the insights!
All the best,
Morse |
|
|
| PauSim |
ITīS NOT THE BATTERIES.
Today I build an AC supply consisting of two 230/15 30 VA (primaries in paralell, secondaries in series) transformers and bridge. Guess what? That distortion still occurs.
So thatīs it. Never mind the finished description of the setup cause breadboards and components and wires are already in the garbage can. Iīm saving batteries to power the kits Iīve just ordered. So letīs put and end to this nightmare (and rejoice at the begining of another) :D
One point I forgot to tell: The distorted sound always happened when the music got more dynamic, particularly when recent recordings were playing, the kind of "normalised", mastered at the software domain. These CDs play at louder levels than earlier ones.
Could it simply be the 1875 clipping? After I assemble the kit, if the problem remains Iīll switch it for the mighty lm3886 Iīve found today.
Well, thanks again for your help guys. We didnīt solve the problem but I learned a lot lately, thanks to your posts.
Bye |
|
|
| Morse |
Hi Paolo;
VERRRRY interesting!
Hmmm, let's see now....that really does sound like clipping with the new info.
Depending on the peak voltage of those transients (which will limit the max output voltage), the amp gain, and the measured power supply voltage, it could be that you just need a beefier chip and more volts in the PS - as you've already suggested. Or it could be that you're overloading the input stage with too high a peak input voltage. Have to look up the max input voltage on the app notes to be sure though.
Good luck and keep us posted. I for one am interested in reading about your battery power experiments (love to do it here, but $$ is tight and I've several cheap 25VCT 50VA trafos on hand....that have worked great on the TDA2030A).
All the best,
Morse |
|
|
|
