http://www.diyaudio.com/forums/software-tools/279762-layouting-free.html
6.39€/Netto/for order of 20!
6.39€/Netto/for order of 20!
Hello hajj and jps1964,
Is simple, not all the people have access to the same resources. In the best part of Europe and EEUU is easy.
As spanish I are in touch with many people in South America, the most part of the diyer's have very limited resources, and need made many times until the mains transformers.
As example you can see some web page as this:
CONSTRUYA SU VIDEOROCKOLA.COM > Toda la información acerca de como construir una video rockola, sus aspectos técnicos, legales y comerciales.
You can see that in all the projects the pcb's are designed for can be self made easily, the autor include versions for serigraphy and for made with iron, no possibility of buy in a pcb company. As example:
http://www.construyasuvideorockola.com/downloads/tda2050.pdf
The reason for I aim a single sided, is for I desire share this project with any people in South America, that cannot buy in a pcb company. So, if someone help for correct the mistakes in my layout have my thanks.
Is simple, not all the people have access to the same resources. In the best part of Europe and EEUU is easy.
As spanish I are in touch with many people in South America, the most part of the diyer's have very limited resources, and need made many times until the mains transformers.
As example you can see some web page as this:
CONSTRUYA SU VIDEOROCKOLA.COM > Toda la información acerca de como construir una video rockola, sus aspectos técnicos, legales y comerciales.
You can see that in all the projects the pcb's are designed for can be self made easily, the autor include versions for serigraphy and for made with iron, no possibility of buy in a pcb company. As example:
http://www.construyasuvideorockola.com/downloads/tda2050.pdf
The reason for I aim a single sided, is for I desire share this project with any people in South America, that cannot buy in a pcb company. So, if someone help for correct the mistakes in my layout have my thanks.
Hello rual_77,
I do understand your concern; I personally live in Lebanon where access to PCB manufacturing is limited at best.
But, in this specific application, routing and layout are very important, so the best way to build it reliably is either to go Point to Point (as Joe himself advised at the beginning of this thread) or go for the 2-layer board designed by JPS1964.
Furthermore, at approx 7 euros/board this is really cheap by any standard. A simple group buy of 20 could be the solution (If you go ahead with it, count me in for 4 boards
)
And btw, a 2-layer board can be done at home, using the same techniques as a single sided board. Might not be as easy, but DIY is all about the challenge, right?
Just my 2 cents
Regards,
Nick
I do understand your concern; I personally live in Lebanon where access to PCB manufacturing is limited at best.
But, in this specific application, routing and layout are very important, so the best way to build it reliably is either to go Point to Point (as Joe himself advised at the beginning of this thread) or go for the 2-layer board designed by JPS1964.
Furthermore, at approx 7 euros/board this is really cheap by any standard. A simple group buy of 20 could be the solution (If you go ahead with it, count me in for 4 boards
)And btw, a 2-layer board can be done at home, using the same techniques as a single sided board. Might not be as easy, but DIY is all about the challenge, right?
Just my 2 cents
Regards,
Nick
I'm interested in trying current drive using multiple LM3886 amps, without passive crossovers, to get acoustic THD & IMD as low as I can.
I see some people are adding a shunt R-C to the amplifier output, where the current is returned through the current sense resistor.
I also thought that the main purpose of using current drive was to control the current in the speaker (rather than the current from the amplifier).
Is the shunt R-C intended to compensate for the increase in acoustic output as the driver impedance rises at HF?
If so this would reduce the benefit of current drive (lower distortion) exactly as the driver moves away from its linear band.
Wouldn't it be better to use quite a small R-C shunt, just to control the amplifier stability at HF, return its current to ground, not through the
current sensing, and then filter the input signal to flatten the frequency response if need be?
The same might be true for an L-C-R shunt tuned to the driver natural frequency (ie, best to shape the signal level FR), but I care less about
that as I intend to use active crossovers to avoid operating my drivers near their natural resonances ... or past where they start to break-up ...
Does this make sense, or have I missed something?
Thanks,
Rod
I see some people are adding a shunt R-C to the amplifier output, where the current is returned through the current sense resistor.
I also thought that the main purpose of using current drive was to control the current in the speaker (rather than the current from the amplifier).
Is the shunt R-C intended to compensate for the increase in acoustic output as the driver impedance rises at HF?
If so this would reduce the benefit of current drive (lower distortion) exactly as the driver moves away from its linear band.
Wouldn't it be better to use quite a small R-C shunt, just to control the amplifier stability at HF, return its current to ground, not through the
current sensing, and then filter the input signal to flatten the frequency response if need be?
The same might be true for an L-C-R shunt tuned to the driver natural frequency (ie, best to shape the signal level FR), but I care less about
that as I intend to use active crossovers to avoid operating my drivers near their natural resonances ... or past where they start to break-up ...
Does this make sense, or have I missed something?
Thanks,
Rod
^ An RC shunt across the load (speaker) is useful for ensuring stability of the current feedback loop and to reduce gain peaking above the audio band where the load is usually highly inductive. I have ended up using 22 - 33 nF which keeps the HF resonance effect far enough from the audio range and also retains good output impedance.
The conventional RC shunt from output to ground serves a different purpose (output stage stability) and may also be helpful in current amps (in addition to the load shunt), but thus far I have not found it necessary.
The conventional RC shunt from output to ground serves a different purpose (output stage stability) and may also be helpful in current amps (in addition to the load shunt), but thus far I have not found it necessary.
Hi Joe, thanks for sharing! Your post and work made me join the forum.
I intend to do stereo current tri-amping with SI1050G (old cheap hybrids) and I found today your schematic, measurements, details, variants and impressions etc.
@Joe, or others, think that it might work as good with hybrids? Any advantage over LMs?
I do not have intentions for full-ranges. Instead a 3-way solution. Quality transformer input then preamplified (current) then buffed(voltage) then your schematics. This ought to solve the overall system&channel&way gain. Apparently this variable gain question is still not answered.
I have three questions more, please help:
1. @Joe@all: Voltage amp + transformer output (windings primary 8ohm : secondary ca 250ohm) + driver 8 ohm would do much worse on sound quality than your solution? Your current source impedance is also ca. 250ohm. The XO+voltage+transformer+narrow band means also all transfer characteristics and components (transformers too) will be easily optimized.
2. @Greebster: Quote "I remember an Analog Devices EE who had a transconductance amp design in one of their Data application books, back 20 or so years ago. Rated for 50w, you were allowed to build two amplifiers without consent for personal use, more contact AD." Really, AD guys not helpful in my experience. Could you please remember the DAB details (volume, years etc)?
3. @all: any literature hints for "current drivers" technical details vs. "voltage drivers" please? I mean loudspeaker (motor) technology mostly, not the rest of details Excuse me, I do not understand quite clear what the definition of "current speakers" is, when we talk small-size 50W full range or not.
I can't wait until having fun with SI1050G's. Thanks for having great fun reading this post.
Cheers!
I intend to do stereo current tri-amping with SI1050G (old cheap hybrids) and I found today your schematic, measurements, details, variants and impressions etc.
@Joe, or others, think that it might work as good with hybrids? Any advantage over LMs?
I do not have intentions for full-ranges. Instead a 3-way solution. Quality transformer input then preamplified (current) then buffed(voltage) then your schematics. This ought to solve the overall system&channel&way gain. Apparently this variable gain question is still not answered.
I have three questions more, please help:
1. @Joe@all: Voltage amp + transformer output (windings primary 8ohm : secondary ca 250ohm) + driver 8 ohm would do much worse on sound quality than your solution? Your current source impedance is also ca. 250ohm. The XO+voltage+transformer+narrow band means also all transfer characteristics and components (transformers too) will be easily optimized.
2. @Greebster: Quote "I remember an Analog Devices EE who had a transconductance amp design in one of their Data application books, back 20 or so years ago. Rated for 50w, you were allowed to build two amplifiers without consent for personal use, more contact AD." Really, AD guys not helpful in my experience. Could you please remember the DAB details (volume, years etc)?
3. @all: any literature hints for "current drivers" technical details vs. "voltage drivers" please? I mean loudspeaker (motor) technology mostly, not the rest of details
Excuse me, I do not understand quite clear what the definition of "current speakers" is, when we talk small-size 50W full range or not.I can't wait until having fun with SI1050G's. Thanks for having great fun reading this post.
Cheers!
^ For this kind of amp, I would choose 15 - 30 ohm; it is not very critical, and the optimum depends a little on the HF inductances expected. Experimenting with real-life loads and a scope can give more insight, but going below 10 ohm can endanger loop stability due to increased feedback factor.
I have three questions more, please help:
1. @Joe@all: Voltage amp + transformer output (windings primary 8ohm : secondary ca 250ohm) + driver 8 ohm would do much worse on sound quality than your solution? Your current source impedance is also ca. 250ohm. The XO+voltage+transformer+narrow band means also all transfer characteristics and components (transformers too) will be easily optimized.
2. @Greebster: Quote "I remember an Analog Devices EE who had a transconductance amp design in one of their Data application books, back 20 or so years ago. Rated for 50w, you were allowed to build two amplifiers without consent for personal use, more contact AD." Really, AD guys not helpful in my experience. Could you please remember the DAB details (volume, years etc)?
3. @all: any literature hints for "current drivers" technical details vs. "voltage drivers" please? I mean loudspeaker (motor) technology mostly, not the rest of details
I can't wait until having fun with SI1050G's. Thanks for having great fun reading this post.
Cheers!
1. If you used a transformer to give you 250 Ohm from a current source, not done the maths, but wouldn't that imply a step-up transformer? I haven't quite got my head around that, but all transformers have a 1:1 or step ratio.
2. Sorry, drawn a blank there.
3. The Elsinore Mk6 is a case in point. It has a crossover and multiple drivers and not fullrange - so it is possible, but it is using an approach that has not found commercial use yet - but we are working on that. In fact, this transconductance amplifier was built to prove to the doubters it was for real - and then decided why not post the details.
But for the average user, speakers without crossovers are easier to current drive. There is also a Nelson Pass PDF article with lots of info to guide.
Cheers, Joe
Hi Joe, thanks for your sharing!
Above:
1. indeed meant as step-up but meantime I realized that this direction has big flaws. This again emphasize the practical elegance of your solution.
2. Well I am waiting for Greebster to remember and hopefully help by sharing details too. If you could contact him in this regards I will greatly appreciate your kindness.
3. I like to find solutions for TIA from voltage-to-current for high power HIFI. In current solid state market one can mostly find products and literature of TIA from current-to-voltage. Even without input stage as current sensing (ultrahigh impedance - tubes) still presents compelling audio advantages vs. the "easy-way" aka. standard voltage world.
In so far, I still do not see the difference between drivers designed for current vs. designed for voltage. At least not at {small, 8ohm, 50W} point. Maybe at 16ohm or 32ohm high efficiency (electrical&acoustic) point, which are neither small or cheap. Same for MK6, due to its high acoustic efficiency, 2xSW+1W+1(M+T) if I read correctly between lines. But... this is not full-range and also not "current specific" design of drivers. Anyway I promise I will read effectively your web site. Great name "Elsinore" btw.!
Success!I hope 3-amping + some preamp tricks + this schematic will be truly refreshing in my home.
Cheers,
Ionmw
For the past couple of months I have been working on optimizing a sealed enclosure for the Markaudio Alpair 10P that would work well with Joe's Trans Amp.
The model so far looks quite promising with a -3db point of 40hz.
I am tempted to post it here so that people can try it out, but first wanted to check with Joe what his opinion about the enclosure was, as there would be a risk for people not liking how the amp would sound simply due to the enclosure not being optimized.
For anyone tempted to run this amp with a run of the mill full range enclosure, my countless models show that one would be underwhelmed by the sound to say the least, and probably could destroy the driver if driven even slightly hard...
Regards,
Nick
The model so far looks quite promising with a -3db point of 40hz.
I am tempted to post it here so that people can try it out, but first wanted to check with Joe what his opinion about the enclosure was, as there would be a risk for people not liking how the amp would sound simply due to the enclosure not being optimized.
For anyone tempted to run this amp with a run of the mill full range enclosure, my countless models show that one would be underwhelmed by the sound to say the least, and probably could destroy the driver if driven even slightly hard...
Regards,
Nick
Last edited:
Hi Nick,
I have a humble question, please share with us from your large experience and understanding why exactly you think:
- that that driver is suitable for current drive
- that if voltage driven would sound bad
- why those other drivers of the "run of the mill" category would sound bad with current source (but not with voltage source... logically implicit)
- also why those will would burn up even if calculated for power and properly current limited (protected) etc.
Thanks and greetings,
Ionmw
Hello Ionmw,
Just a quick correction, the driver I am modeling the enclosure around is the Alpair 10.3
To answer your questions:
- The driver itself is not the best to be current driven, and actually I am yet to find a full-range driver in production that fits the bill. It just happens to be a driver I am familiar with that has a low Qms (and according to Planet10, even this low Qms is still twice as large as one would like it to be)
My understanding of the idea behind a low Qms driver is that current driving means that there will be no electrical damping of the driver around Fs, and therefore one would have to rely on mechanical damping.
- I never said the driver would sound bad if voltage driven.
- My comment was about the enclosure and not about the driver. Designing an enclosure for current drive is very different than for voltage drive.
If one were to drive a voltage-drive optimized enclosure with a current drive amp, the result would be a big hump in the freq response around Fs (remember that a current drive amp will deliver more voltage into a rising impedance)
- For the same reason as above, driver excursion around Fs would be wild and the risk of damaging the suspension and/or cone is a real one.
Hope this answers your questions.
Regards,
Nick
Ideally you want to find a driver with a Qms <1.0 if you want to use it with a current amp unaided.
Common knowledge suggests that box Q is always higher than driver Q but Esa descibes a speaker where he lowers the box Q.
Joe has also descibed a circuit that flattens the impedance.
Aperiodic TLs & boxes can be used to flatten the impedance curve.
dave
I am not fully involved in fullrange and indeed the Elsinores are a 'system' that is designed to be current friendly.
So asking those into fullrange drivers, what is desirable for current drive? I would invite opinions please, from others.
But low Qm has to be an adavantage and low inductance. Anybody else would add to that?
Cheers, Joe
You beat me to it.
Indeed on all three. I have Esa's book - anyone else here?
Esa has his 'style' that may not rub some people well, but I do recommend it.
A tuned LCR can flatten out the response, matching frequency and total Q. Let's say you adjust volume to be a 2nd order Butterworth when driven from a voltage source, add LCR and now it will always be a 2nd order Butterworth no matter what the drive impedance is, hence current compatible. The response will follow the predicted Thevenin's theorem - and if the response is that of a 2nd order Butterworth, then it has the Q of a 2nd order Butterworth. The response has been divorced from the amplifier, the alignment determined by only Vb.
Some years ago I mentioned this to some local designers down here in Australia and they were aghast. That it works and is so simple. But I raised the question "since Qm is unchanged in situ, what happens to the electrical Q?" Blank faces. But these guys knew me, but when I mentioned this on a few forums, I was accused of lying.
Deja vu?
But... again... anybody can try it and see that it works.
Also great for nfb single-ended amplifiers. And you don't need a very low Qm driver. Works best for sealed boxes, but if vented, then tune for the higher peak and flatten that and in most cases the lower peak will also be suppressed somewhat and lowered in frequency. Also works for line loaded drivers, TLS.
It's best to have software to come up with the right LCR values. I use SoundEasy after importing the Z plot.
Last edited:
Bought it and read it a couple of years ago.
I had been looking around ever since for an easy to build current drive amp to use with a full range enclosure.
The project shown in the book is too complex and more importantly, being from the single-driver school, I usually prefer minimal signal shaping, and this is where your Class D amp comes into play.
Care to elaborate a bit on that point?
Thanks
Nick
With vented alignment, something to model it is really necessary. The LC values are chosen for the frequency of the upper peak, say 70-80Hz. The R value controls the level/degree of flattening. The key here is the find the combination of LC that does not change with frequency, but changes the Q. Once the right Q is found, then adjust R. You may have to go forward and and backwards to get the flattest result. If the lower peak was 30 Hertz approx and say Z = 20 Ohm, then the Q of your LCR will overlap and push it down to maybe 20-25Hz and at the frequency the Z will be lower than it is at the original 30Hz.
I use SoundEasy, measure the Z plot and import, then play around with the values that gets the best result. You ideally should end up with a single peak way below Fb. There will be other software equally capable. The only way to demonstrate completely would be a video.
Example Elsinore Mk6:
There are no commercial speakers that gets anywhere close to this kind of result.
Last edited: