Survey on class D project

Hey you great Guys

We are a bunch of boys who developed a new class D amplifier module originally intended for personal use.
But due to pretty high costs, we thought maybe we should make a batch, and share the module with others to lower the costs pr. module.

Anyways here are a few high lights of the module:

It is a self-oscillating class D amplifier with brand new twin feedback loop technology.
The output MOSFETs are IRF Direct FETs smd mount types with very low induction.
Output power is 1 X 450 Watts @ 4 Ohm @ +- 60V supply.
Calculated Zout @ 1KHz = 9µOhms
Gain 25 dB
Noise 25 µV
Phase shift from input to output 0,0 Dgr throughout the audioband.

What I would like to know is:
Would it make sense to combine a kit with i.e. a complete powersupply?
What kind of powersupply would be the preferred one:
A somewhat costly custom switch mode power supply?
A standard cheaper SMPS form third party?
A linear supply with custom toroids and PCB for capacitance and regulators for +-12V?
No supply at all?

The input impedance of class D amplifiers are often pretty low, as the impedance is a choise between noise and impedance.
Thus a buffer can be necessary to ease the load on a preamplifier or a DAC.
We´ve thought of three different types, which could be options:

An unbalanced push pull unity gain buffer with discrete technology and no negative feedback, and fully DC coupled operated with servo circuits.
A Balanced buffer with op-amps OPA1612
A buffer with footprints for cross overs, to fit an active speaker.
You can then choose filtertype, slope and frequency on your own, formulas provided though.

The buffer PCB will fit onto the connectors on the amplifier board, so no soldering is required.

The THD measurement below is done under less than ideal conditions i.e with a lab supply, not capable of any more power than shown, and long leads etc.
The module on the pictures is a hand soldered prototype, the final layout will not be changed very much though.

We will very much appreciate your opinion, ideas or thoughts.
And of course questions are most welcome.
 

Attachments

  • THD+N bigFOOT MKII.pdf
    103.6 KB · Views: 108
  • bigfoot MK2,2 .jpg
    bigfoot MK2,2 .jpg
    305.4 KB · Views: 337
  • bigfoot MK2.jpg
    bigfoot MK2.jpg
    127.8 KB · Views: 356
Last edited:
Hi Kurt,

Good initiative! You state 450W in 8 Ohm with a +/-60V power supply. +/-60V should allow 450W in 4 Ohm only (SE configuration), am I right? If bridged, it would be some 900W in 8 Ohm.

Oh damned you are sharp :)
Sorry for the mistake.
Of course it is at 4 Ohms.
Anyways it is also fully stable at 2 Ohms, and it is a half bridge design, which makes it bridgeable.

Do you have any considerations regarding powersupply?
 
Kurt

This sounds like something I am interested in. I do not currently have a class D amp in the inventory. As far as power supply, it is possible for anyone here to DIY a linear supply or go to ebay and buy a cheap +- switching supply. A premium performance switching supply designed to match this amplifier seems to me like a good choice as an offering for those who don’t want to go for one of the other 2 choices. An add on buffer board that offers no solder connection to the main board is a DIY dream come true! Thank you for bringing this to the forum.
 
Kurt

This sounds like something I am interested in. I do not currently have a class D amp in the inventory. As far as power supply, it is possible for anyone here to DIY a linear supply or go to ebay and buy a cheap +- switching supply. A premium performance switching supply designed to match this amplifier seems to me like a good choice as an offering for those who don’t want to go for one of the other 2 choices. An add on buffer board that offers no solder connection to the main board is a DIY dream come true! Thank you for bringing this to the forum.

Hi Bfpca
Thank you for your considerations.
I think you are right about the powersupply.
We´ve already built a premium SMPS intended for driving the class D amp.
The problem with really good SMPS is costs. It is impossible to work arround custom wound transformers i.e.
But if cost is no object, we could provide a single stage BCM resonant converter with active PFC, which is as silent as the grave both forwards and backwards.

The idea with the buffer designs is, that they can be interchangeable, so that the module will fit most purposes.
Right now we are doing final eksperiments with the bufferdesign, so not everything is yet decided. I.e we will probably use onboard shunt regulators for the buffer supply in the discrete design.
The design is non feedback with DC servo´s, and it has quite crazy data.
The op-amp based ones will be a bit more normal designs, except the one with filter footprints for different filtertypes, slopes and orders, if any interest.

Anyways the THD measurement was made under less than ideal conditions.
I.e. with a lab supply, it will be a lot better when final measurements are made. We will save that for the datasheet, when ready.
 
Another high quality amp board for DIYS (if connecting some wires to a module is DIYS), would always be welcome. I hope you have success with it.

I thing you have looked into the market and made your calculations. You are sure aware that there are different categories of competing amps are around.

One is the junk class where people with unknown education and physical background have copied some Western design, saved cost where possible and are producing it in various Chinese backyards. Some of it, like the Ljim stuff, sometimes even works without hurting your ears or blowing your speakers in the first minute.
This forum has had quite some influence on these designs by improving them... Clever, if you find stupid, educated people that buy your unfinished stuff and make it work, so you can produce a better MK2 version. Then you sell the amp you where unable to do by your self with your name.
These guys will be happy to copy your design, as they will buy the first module you make available. Be sure some “Bill” from GB or “Hans” from Germany will sent it to Shenzen copy center in no time.
It is not to make profit if they copy you, but their way of showing you how much they value your work. Other culture...
As their home country even pays for the shipping cost, you are no competition for them.
Any trick you try to make copying impossible is known by them. Not providing a Schematic helps nothing, they can't even read it, they just copy 1:1. Copy protection only hurts your customers, makes after sales service a nightmare and gives you a bad name.

Then, there is the European “Hype” class, like ICE, Abletec, Pascal, Hypex etc. If still in business, they have their main focus on OEM supply. They prohibit end user sales or sell the same stuff, the OEM gets for 20 Cent, to the DIYS market for a Dollar. They can undercut your best price by 50% and still make more money than you hope to earn. They have so many patents on any ridiculous detail, I don't think you could not even build a light bulb without being sued. If your amp is worth thinking about.

Beside from this, I may have some practical thoughts about such an amp.

If you ask for “extra stuff” people would like, you will end up with full blown AVR receiver including HD- LCD screen and beer cooler.

One want´s a symmetric input, the next an x-over. For the next guy a DSP is a must be, Bluetooth, integrated SMPS etc.
In the end anyone will like it, but no one pay the final price.
Keep it simple, anything you can offer as add on will be cheaper somewhere else. So don't do plug in stuff.

If you ask me, Do a really good amp and offer an optional SMPS any idiot can wire up. Plug&Play. If possible, make the input with a buffer that takes sym and asym. That´s it. No more. Two parts. Any additional PCB gets too expensive.
An Online tutorial video is the best you can do at low to no cost.
People do not like to read any more. The skilled guys that can wire up complicated stuff are found less often than you think.
DIYS people do make mistakes you can't even dream of. Make it idiot proof high end. Include anything one can need, so no special connectors as side order.

I think there is a trend coming, where some people retrofit old amps cases with new internals, as energy wasting stuff becomes a No-No in many countries. Maybe hop on that train with one leg.

Good luck!
 
Another high quality amp board for DIYS (if connecting some wires to a module is DIYS), would always be welcome. I hope you have success with it.

I thing you have looked into the market and made your calculations. You are sure aware that there are different categories of competing amps are around.

One is the junk class where people with unknown education and physical background have copied some Western design, saved cost where possible and are producing it in various Chinese backyards. Some of it, like the Ljim stuff, sometimes even works without hurting your ears or blowing your speakers in the first minute.
This forum has had quite some influence on these designs by improving them... Clever, if you find stupid, educated people that buy your unfinished stuff and make it work, so you can produce a better MK2 version. Then you sell the amp you where unable to do by your self with your name.
These guys will be happy to copy your design, as they will buy the first module you make available. Be sure some “Bill” from GB or “Hans” from Germany will sent it to Shenzen copy center in no time.
It is not to make profit if they copy you, but their way of showing you how much they value your work. Other culture...
As their home country even pays for the shipping cost, you are no competition for them.
Any trick you try to make copying impossible is known by them. Not providing a Schematic helps nothing, they can't even read it, they just copy 1:1. Copy protection only hurts your customers, makes after sales service a nightmare and gives you a bad name.

Then, there is the European “Hype” class, like ICE, Abletec, Pascal, Hypex etc. If still in business, they have their main focus on OEM supply. They prohibit end user sales or sell the same stuff, the OEM gets for 20 Cent, to the DIYS market for a Dollar. They can undercut your best price by 50% and still make more money than you hope to earn. They have so many patents on any ridiculous detail, I don't think you could not even build a light bulb without being sued. If your amp is worth thinking about.

Beside from this, I may have some practical thoughts about such an amp.

If you ask for “extra stuff” people would like, you will end up with full blown AVR receiver including HD- LCD screen and beer cooler.

One want´s a symmetric input, the next an x-over. For the next guy a DSP is a must be, Bluetooth, integrated SMPS etc.
In the end anyone will like it, but no one pay the final price.
Keep it simple, anything you can offer as add on will be cheaper somewhere else. So don't do plug in stuff.

If you ask me, Do a really good amp and offer an optional SMPS any idiot can wire up. Plug&Play. If possible, make the input with a buffer that takes sym and asym. That´s it. No more. Two parts. Any additional PCB gets too expensive.
An Online tutorial video is the best you can do at low to no cost.
People do not like to read any more. The skilled guys that can wire up complicated stuff are found less often than you think.
DIYS people do make mistakes you can't even dream of. Make it idiot proof high end. Include anything one can need, so no special connectors as side order.

I think there is a trend coming, where some people retrofit old amps cases with new internals, as energy wasting stuff becomes a No-No in many countries. Maybe hop on that train with one leg.

Good luck!

Very interesting thoughts mr. Turbowatch2.
Thank you for your considerations.

We know that fixing a class D amp and a SMPS in a cabinet is maybe not the wildest DIY project one can imagine, but 4 layer PCB Layout, handsoldering around 200 0805 components on a PCB not much bigger than a credit card, would hardly end with succes for anyone apart from experts, so that is no option.

We know about the way they do business in China, and they are of no interest to us.
Our aim is to match the european and mostly Danish designs, as three of the best and/or largest ones are Danish.
Apart from them, we wish to make a DIY module available for DIY projects.
What we might do later on, we don´t know yet, but the design of this amp is IMO way too good just to be kept a secret.
We are aware of patents, almost all of them were kept by ICEPower, and has already expired.

I understand that you think, that if you reach a hand to the devil, he will tear of your arm :D
You might be right on that mater.
The reason we intend to provide buffers, is because of the somewhat lowish input impedance, which could cause trouble upstream. But some preamps might cope with it anyway, and therefor we think, that use without buffer should be offered.
I also understand that you think a good SMPS with plug and play connectors would be the best to offer.

Thank you for your response
 
Last edited:
Turbowatch2 basically said everything.

I myself and I think possibly others who like power wasting class A with lots of heat. For those of us making flea amps or full range speaker cabinets and even the high powered ones are lacking a good controlling amplifier to insert a subwoofer for a multiple amp set up.
As of yet I have no experience building class D amplifiers to fill that specific purpose for subs as for power supplies a lot of general market ones that are available some from Germany that might fill the niche. keep it simple.
If your design is really that good and it proves to be an excellent amplifier for subs I will show it out for display on next year‘s Amp Camp Amp 2020 with my setup.
 

29285

Disabled Account
2010-01-17 9:20 pm
I totally agree with Turbowatchs statements and would encourage you to keep things as simple as possible for the DIYer. The only extravaganca to consider is a tiny AFE (audio front end) as interconnect between audio source and amp input. This can be a hi-Z symmetrical input with XLR connector, or a limiter, or some crossover etc ppp. The idea is a variety of inexpensive 2-layer AFEs to taylor customers requirements without changing the 4-layer power modul. The AFE with input jack(s) might be mounted to the front panel while the power module is somewhere else. The connection will be some unshielded ribbon cable with IPC connectors providing power supply, gnd and symmetrical line level.
just my 2c
 
Turbowatch2 basically said everything.

I myself and I think possibly others who like power wasting class A with lots of heat. For those of us making flea amps or full range speaker cabinets and even the high powered ones are lacking a good controlling amplifier to insert a subwoofer for a multiple amp set up.
As of yet I have no experience building class D amplifiers to fill that specific purpose for subs as for power supplies a lot of general market ones that are available some from Germany that might fill the niche. keep it simple.
If your design is really that good and it proves to be an excellent amplifier for subs I will show it out for display on next year‘s Amp Camp Amp 2020 with my setup.

Hi ThermalAlchemy
Thanks for your input.
If 450 Watts @ 4 Ohms is sufficient, the module will be great for subs, as its Zout is very low. We did not yet measure the Zout, until now it is only calculated. We have to make a Kelvin measurement to find the real life value, but it will still be ludicruos low.
And as the amp has no phase shift what so ever through out the audioband, it will be just as good for highs than it is for lows.
The protection circuits will allow up to 30A current, and also protect against DC output, as the amp is fully DC coupled, and the buffer utilises very slow servo circuits. You can call it "subproof" If I may say so :)

I understand you too would like the easy solution with a matching premium SMPS and the relevant buffers.
I don´t really know any german vendors of premium class d amplifiers.
The ones that we try to match are ICEPower, Hypex and Purify, of which none are german.
 
I totally agree with Turbowatchs statements and would encourage you to keep things as simple as possible for the DIYer. The only extravaganca to consider is a tiny AFE (audio front end) as interconnect between audio source and amp input. This can be a hi-Z symmetrical input with XLR connector, or a limiter, or some crossover etc ppp. The idea is a variety of inexpensive 2-layer AFEs to taylor customers requirements without changing the 4-layer power modul. The AFE with input jack(s) might be mounted to the front panel while the power module is somewhere else. The connection will be some unshielded ribbon cable with IPC connectors providing power supply, gnd and symmetrical line level.
just my 2c

Thank´s a lot for your ideas.
The AFE´s we´ve considered, and for the time being are prototyping, are small non solder PCB´s that fits into the boards connections, and then have its own set of input connectors.
You can have a look at the PCB design in the picture. The buffer is on the left side.
Yet it might be looking different when devellopment is completed
 

Attachments

  • 73201496_2303440626635600_465005981605036032_n.jpg
    73201496_2303440626635600_465005981605036032_n.jpg
    134.2 KB · Views: 179
Last edited:
Tell me what do you think of these products I have not purchased them yet I was looking at their switch mode power supplies for a clean power source. But they also make amplifiers
DIYclassd.com

We have used Hypex nCore 500 OEM modules as one of our references, and it has very good specs. As DIY´er you can only buy nCore 400 modules or the cheaper UcD amplifiers.
The powersupplies are fine, but unregulated.

Be aware that the OEM´s only have arround 12-14 dB voltagegain, and need buffering.
The NC400 has arround 25 dB voltagegain and Zin @ 100 KOhms
They generally measure as they say.

The sound is often discussed, some find them a bit bright and lean, others as no nonsense.
What could be mentioned as a critical comment would be the quality of the components i.e. electrolyttic capasitors. I´d prefer more reputated vendors known for reliability, but maybe I´m old fashioned.
 

29285

Disabled Account
2010-01-17 9:20 pm
Nice to see you already had a similar idea.
Concerning the quality of hypex modules, I think they are on their way earning money with OEM modules and now the bean-counters enter the scene. That's business as usual.
For better quality and amps that deserve the hi-end attribute DIY seems the way to go:p
 
Nice to see you already had a similar idea.
Concerning the quality of hypex modules, I think they are on their way earning money with OEM modules and now the bean-counters enter the scene. That's business as usual.
For better quality and amps that deserve the hi-end attribute DIY seems the way to go:p

In this new class D module quality was paramount for every component.
I.e. electrolytics are either Nippon Chemicon or Panasonic with lowest possible ESR, and 105 Dgr C temperature marking marking. The MOSFETs are ultra low induction direct FETs from IFR, the modulation is taken care of by an OPA1612, and the design itself is not really putting stress on the components, which is one of the new things in this concept. There is really not components anywhere, which are near thier limits regarding neither current/voltage nor speed. The latter being the most common problem.
Thus many years of operation can be expected.
 
Hi Kurt,

Thanks for providing fast replies to the different postings. It shows that the efforts of the posters are appreciated.
Sorry, I cannot join the project myself. I already have so many amplifiers that my wife, with raised eye-browses, asks: “What are we going to use all this for?”

On the power supply:

The amplifier will be used with 4 to 8 Ohm speakers (2 Ohm is less realistic). Using +/-60V supply, you have 4 different power levels – 8 Ohm/SE coupling/225W; 8 Ohm/BTL coupling/900W; 4 Ohm/SE coupling/450W; 4 Ohm/BTL coupling/1800W! Thus, the users will have power supply demands ranging from at least 225W (per channel) to 1800W! (per channel). You hardly want to bother with so diverse individual needs. Some may even prefer a lower supply voltage and you have even more types of power supplies.

On power supplies, audiophiles seem quite divided on traditional net-transformer supplies or SMPS. For the most delicate audiophile listening, perhaps the traditional power supply is preferred while for high power use, an SMPS is often the most practical.

SMPS: Particularly suited for less precise voltages and operation within a more limited power range. Less suited when very stable voltages are demanded and the power consumption may be both very low and high. In order to handle very low power consumption, an SMPS often uses a cycle-skipping or burst-mode, which may affect audio quality. While building even an 1800W SMPS is possible, achieving very fast load response over a large loading range is far from trivial. You may end up spending at least as much effort on designing such an fast-responding SMPS as you used on designing the amplifier. A very powerful “one-fits-all” will hardly be feasible as it for most uses will be operating at the low end of its range with problems near idle loading.
I would refrain from designing your own SMPS and look for an existing SMPS, made for audio amplifiers, to suggest for use with the amplifier if an SMPS is preferred.

Traditional net-transformer based power supply: Much easier to adapt to individual wishes for power-level and eventual supply voltage. If you suggest the amplifier users a good 3rd-party board substantially with a rectifier bridge and generous high-quality decoupling capacitors, you can let each user choose his own transformer to be used with it.

Further features:

A balanced input and band-pass-filter front-end to be individually adapted as needed will be an appreciated feature. But, as many components need to be chosen by the users and should be of the “through-hole” type, inclusion of such functionality will hamper your compact amplifier layout. If you could add such functionality on a separate board fitted to be connected to the amplifier board, I believe it will be appreciated.

:up:
 
Last edited:
Hi Kurt,

Thanks for providing fast replies to the different postings. It shows that the efforts of the posters are appreciated.
Sorry, I cannot join the project myself. I already have so many amplifiers that my wife, with raised eye-browses, asks: “What are we going to use all this for?”

On the power supply:

The amplifier will be used with 4 to 8 Ohm speakers (2 Ohm is less realistic). Using +/-60V supply, you have 4 different power levels – 8 Ohm/SE coupling/225W; 8 Ohm/BTL coupling/900W; 4 Ohm/SE coupling/450W; 4 Ohm/BTL coupling/1800W! Thus, the users will have power supply demands ranging from at least 225W (per channel) to 1800W (per channel). You hardly want to bother with so diverse individual needs. Some may even prefer a lower supply voltage and you have even more types of power supplies.

On power supplies, audiophiles seem quite divided on traditional net-transformer supplies or SMPS. For the most delicate audiophile listening, perhaps the traditional power supply is preferred while for high power use, an SMPS is often the most practical.

SMPS: Particularly suited for less precise voltages and operation within a more limited power range. Less suited when very stable voltages are demanded and the power consumption may be both very low and high. In order to handle very low power consumption, an SMPS often uses a cycle-skipping or burst-mode, which may affect audio quality. While building even an 1800W SMPS is possible, achieving very fast load response over a large loading range is far from trivial. You may end up spending at least as much effort on designing such an fast-responding SMPS as you used on designing the amplifier. A very powerful “one-fits-all” will hardly be feasible as it for most uses will be operating at the low end of its range with problems near idle loading.
I would refrain from designing your own SMPS and look for an existing SMPS, made for audio amplifiers, to suggest for use with the amplifier if an SMPS is preferred.

Traditional net-transformer based power supply: Much easier to adapt to individual wishes for power-level and eventual supply voltage. If you suggest the amplifier users a good 3rd-party board substantially with a rectifier bridge and generous high-quality decoupling capacitors, you can let each user choose his own transformer to be used with it.

Further features:

A balanced input and band-pass-filter front-end to be individually adapted as needed will be an appreciated feature. But, as many components need to be chosen by the users and should be of the “through-hole” type, inclusion of such functionality will hamper your compact amplifier layout. If you could add such functionality on a separate board fitted to be connected to the amplifier board, I believe it will be appreciated.

:up:

Hi FauxFrench
Both yours and all of the other users posts are very much appreciated, and they are very important to us.
You are right about the difference in power delivery for the module, and I guess it will always be like that.
But one of the points with SMPS is, that it will not get better with size.
Either it works perfectly or it shuts down. So in the real world it is not necessary to dimension SMPS for full sinus waves at full power.
That is only relevant for continuos power measuring.
Music will never call upon that much power, and as our SMPS is fully regulated, it will hardly ever get into any trouble with music of any kind, unless the load is rediculous.
And this kan be done with just 600 Watts smps for TWO channels.
But not many audiophiles beleive in stuf like that, unregarded that a welldesigned 600 Watt smps can easily deliver peaks 3 times its marking, and that music never needs power exceeding 1/8 th of a sinus.
Anyways our intention is to make a smps meeting reasonable power delivery and absolute top specs on noise and ripple.
One can then feel free to use one or two of them, that should be sufficent for any use.
We´ve been looking for third party smps, but none of good quality was found besides Hypex, and they are unregulated, which is rather disappointing.

I agree to the linear supplies, people will probably like very large ones, and I think there is a lot of different viewpoints on rectifiers and so on.
So maybe we can recommend something at least.

Regarding the buffers, they are all smd design, but it should be possible to solder filter components by hand. Maybe we can choose a bit larger component size for the filters.
Anyways THT components availability is decreasing rapidly, so we find smd a better choice in the long run.

Best regards
 

29285

Disabled Account
2010-01-17 9:20 pm
Some remarks on power supply. I designed on 100x100mm a LLC-converter that delivers 400-500W continuously without forced air cooling. With synchronous rectification the efficiency is close to 97%. Clamping diodes make it short circuit proof, thus there is no need for a softstart or integrated current limit. It works unregulated with fixed frequency set to series resonant frequency yielding optimal working condition over the entire load range. No pulse skippin on low load, and ZVS over the entire load range.
Thus output impedance is minimal and voltage fluctuation vs load as well, outperforming similar toroidal xformer supplies.
All in all this is a quite simple smps with minor stress on all components. A direct replacement for a toroid supply with improved performance.
 
Some remarks on power supply. I designed on 100x100mm a LLC-converter that delivers 400-500W continuously without forced air cooling. With synchronous rectification the efficiency is close to 97%. Clamping diodes make it short circuit proof, thus there is no need for a softstart or integrated current limit. It works unregulated with fixed frequency set to series resonant frequency yielding optimal working condition over the entire load range. No pulse skippin on low load, and ZVS over the entire load range.
Thus output impedance is minimal and voltage fluctuation vs load as well, outperforming similar toroidal xformer supplies.
All in all this is a quite simple smps with minor stress on all components. A direct replacement for a toroid supply with improved performance.

I agree that often smps can be the very best choise for audio amplifiers, amongst others for the reasons you mentioned.
They are very stable, and some designs are also very quiet, and can be recommended for linear power amps as well as class D.