• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Amplifier Modules and PCBs For Sale

Status
Not open for further replies.
D

Deleted member 148505

A year ago, I was scouring for an amplifier for my subwoofer project. So I've read some of the class d threads here, and learned many ideas because of it. I've come across a sigma-delta class d amplifier, created my own layout, built it, and I liked the sound of the bass, better than most class AB Ive tried., but the amp was easily busted because I accidentally short circuited the outputs.

Since I know basic electronics and can make PCB layouts. I thought of a design that combines most of the desirable features of the existing ones freely distributed here.

One of the easily implementable and forgiving design I found on the forum are the derivatives of IRAUDAMP1, a self oscillating sigma-delta class-D, implementation of International rectifier. There are also loads of documentation/application notes that can be freely downloaded on the internet that's why I chose that for the PCB layout of my subwoofer amplifier project.

Here are some of the design considerations for my amplifier project.

- The first goal of my design is that the driver IC must be easily obtainable and cheap. Must be cheaper than IRS2092.

- Must be all through hole components and low parts count. All through hole so that the amplifier will be easily assembled and repaired by an amateur.

- Another goal for the design is that the power supply for the driver chip must be integrated on the PCB. I chose the LM350 regulator as the power supply because it can easily provide 2A of current for the mosfet driver. I included the circuitry in the PCB such that only an external AC voltage will be needed. Also when there's an external supply for the driver, the amplifier will accept wide voltage range for the rail supply. The driver supply should be mounted on a heatsink too. Heatsink is calculated based on dissipation.

- Another goal for the design is to have an overcurrent protection built in. One of the easy implementation I found is the low side current sensing using NE555 and LM311.

- Single sided PCB. I created a single sided pcb for my prototype. But my latest design uses double layer to reduce noise.

- No built in speaker protection with relay on board because based on my experience, it is difficult to replace a busted relay without (slightly) damaging the pcb board.

- Paralleled outputs. One of the original goal is to have parallel output. I tried it and it worked on my prototype but the driver IC and the regulator IC is heating up like hell. So based on actual use and my calculations, the reliability will be compromised, if there's no buffer for the gate driver. So my final design uses only 1 pair of output mosfets.

Here are some of the application notes that helped me in building this amp.

Reference Design:
IRAUDAMP1 RevB

IR Class D Tutorials:
Class D Audio Amplifier Design
Designing Practical High Performance Class D Audio Amplifier

IR Design Tips:
DT04-4 Using Monolithic High Voltage Gate Drivers
DT97-3 Managing Transients in Control IC Driven Power Stages
DT99-7 Alleviating High Side Latch on Problem at Power Up

IR Application Notes:
AN-1071 - Class D Audio Amplifier Basics
AN-1084 - Power MOSFET Basics
AN-978 - HV Floating MOS-Gate Driver ICs
AN-1070 - Class D Audio Amplifier Performance Relationship to MOSFET Parameters
AN-1135 - PCB Layout with IR Class D Audio Gate Drivers

Here's the picture of my prototype. The bass is very good. But in full range, sound is just ok. Also when it is connected directly to my laptop, there's an audible hiss, but when I used other sources, there are no hiss. That's why in my final design, I used double layer for the PCB. And it solved the hiss problem :) .

The amp is beside by receiver. Interference on FM radio stations is also not a problem. Better than a mono 250W class D I bought on some ebay store, the one I bought makes the FM radio of my receiver unusable. But I haven't tried it on AM though.

One of the problems I encountered was the heating of driver IC and regulator IC. So on my succeeding design I provided ample space for the heatsink for the voltage regulator. I also avoided paralleling of mosfets.

I only have a 20mhz scope and ordinary probes. As you can see there is some minor undershoot.
 

Attachments

  • IMG_3527.jpg
    IMG_3527.jpg
    383.8 KB · Views: 5,412
  • IMG_3465.jpg
    IMG_3465.jpg
    423 KB · Views: 4,219
  • IMG_3537.jpg
    IMG_3537.jpg
    432 KB · Views: 3,808
  • IMG_3466.jpg
    IMG_3466.jpg
    347 KB · Views: 3,574
Last edited by a moderator:
D

Deleted member 148505

Here's my second iteration of the design, I've added an SMD buffer totem pole to the gate drive to get the burden of driving the mosfets instead of the IR2110 IC.

But after reading an application note about the consequences of paralleling mosfets, and since I have a limited knowledge and testing tools to measure the amp would be, I realized that bridging is a better solution to get more power that's why I discarded my 2 pairs design.

Also the buffer totem pole SMDs is hard to solder and the price is relatively high, and is not easily obtainable.
 

Attachments

  • 2nd iter.PNG
    2nd iter.PNG
    72.7 KB · Views: 4,191
Last edited by a moderator:
D

Deleted member 148505

This is the final design of the layout of my class D amplifier. As you can see, the top copper is used as shielding planes. Also I avoided to put the shielding plane on the noisy high side gate drive. Just to be sure, because I've read that some of our members here didn't get good results with filling the top layer with copper plane. I got good result with this one.

The hiss was totally removed by the layout. Also interference to FM radio at 1ft is negligible.

I also made the mosfet TO247 and TO220 compatible.
I used HEF4069 as hex inverter, and IR2110 as gate driver. :)
The amp was designed to theoretically handle 20A output.
 

Attachments

  • Functional Description.PNG
    Functional Description.PNG
    40.2 KB · Views: 1,725
  • Functional Allocation.jpg
    Functional Allocation.jpg
    290.7 KB · Views: 2,211
  • diag.PNG
    diag.PNG
    56.9 KB · Views: 1,126
Last edited by a moderator:
D

Deleted member 148505

I don't have a stiff power supply at hand yet so I didn't push it hard.

I'll publish my power measurements after I received my power supply. (ETA about 2 weeks) :)

I'll gonna use +-75V rails.
Target output current will be 16A into 2 ohms and 11A into 4 ohms (500W RMS)

Actual results might be better :)
 
D

Deleted member 148505

@stewin
Theoretically you can use higher rails for this amp if you put 250V mosfets (IRFB4229) and replace the value of resistor for the regulator of TL071. You also have to use higher voltage value for the bus capacitors.
 
D

Deleted member 148505

Nice work sir .. it is looks like drtex irs 1500 !!

Yes, but much better in terms of layout, features and performance. They are look alike if you see them from afar. If you examine them closely, they are very much different. :D

I also added extra measures (circuit and layout) to avoid the possibility of high side latching during startup and transients. And strictly followed layout guidelines as much as possible :)

Here's the sample vid of my amp just after I finished it assembling.
JLAmp1000D Subwoofer Amp Running Without Heatsink - YouTube
 
jlester87 hello, I think it's great PCB design, a few months ago I did a very similar design and found problems with noise that you mentioned, as was working with a layer, the quickest solution was to increase the switching frequency at 450KHz but that has its counterparts.

greetings ...
 
D

Deleted member 148505

Back then I didn't increase the switching frequency further to 300kHz because I am careful not to overheat the IR2110 as I was using IRFB4127 mosfets which has a Qg value 150nc.

With my current layout there are no noise problems with my laptop as a source. EMI is good too. And undershoot is somehow minimized.

Can you show the pics of your amp?
 
D

Deleted member 148505

I don't have a stiff power supply at hand yet so I didn't push it hard.

I'll publish my power measurements after I received my power supply. (ETA about 2 weeks) :)

I'll gonna use +-75V rails.
Target output current will be 16A into 2 ohms and 11A into 4 ohms (500W RMS)

Actual results might be better :)

My 62-0-62VAC @240VAC (15A) transformer has arrived. Filter caps are 6800uf 100V rated.
Dummy load is configurable to 8/4/2 ohms. I'll submerge it in water

I'll make power measurements next weekend :)
 

Attachments

  • 9224671180_fede069885_b.jpg
    9224671180_fede069885_b.jpg
    304.2 KB · Views: 2,268
  • 9224697350_0a56c7dcd1_b.jpg
    9224697350_0a56c7dcd1_b.jpg
    358.5 KB · Views: 2,146
Back then I didn't increase the switching frequency further to 300kHz because I am careful not to overheat the IR2110 as I was using IRFB4127 mosfets which has a Qg value 150nc.

With my current layout there are no noise problems with my laptop as a source. EMI is good too. And undershoot is somehow minimized.

Can you show the pics of your amp?

Hello jlester87, the image shows the prototype that mountain, I have no noise problem with a linear source despite this fact in one layer without display earth.

when connected to a SMPS, I hear a high frequency noise, to mitigate a bit, increase the frequency of 250kHz to 400kHz, but this is not desirable because it heats up the IR2110.

I can you please make 2 measurements with the oscilloscope, to make a comparison of forms of waves that do not seem quite right?

regards...
 

Attachments

  • Captura 3.JPG
    Captura 3.JPG
    64 KB · Views: 1,001
  • Captura 2.JPG
    Captura 2.JPG
    49.5 KB · Views: 931
D

Deleted member 148505

Looks interesting,will you post an pcb layout?,can this run at 20vdc x2?

PCB layout is not for sharing, and it's double sided, so it's difficult to etch. I will release 2 pairs version of this and will mass produce the PCB so it will be cheap, around 12USD per board. (excluding components)

Advantage of using an external supply for the driver IC is that it will work at wide range of power supply voltage, I tested this to work from +-35 to +-75VDC without changing any resistors ;)

You should use single(independent) 12 to 14VAC 2A EI core for the driver IC supply, external 14 to 16VDC should work too.

Hello jlester87, the image shows the prototype that mountain, I have no noise problem with a linear source despite this fact in one layer without display earth.

when connected to a SMPS, I hear a high frequency noise, to mitigate a bit, increase the frequency of 250kHz to 400kHz, but this is not desirable because it heats up the IR2110.

I can you please make 2 measurements with the oscilloscope, to make a comparison of forms of waves that do not seem quite right?

regards...
I haven't tested this with SMPS yet, I don't have any. I'm using linear supply and noise is negligible when you put your ear on the tweeter. Sound is very good too.. approaching the SQ of most TDA8950/IRS2092 class D amps sold on ebay. (yes i already heard them)

Can you specify what measurement you would like to see? I can't understand your question.

I suggest that you should add additional filtering after your SMPS, material 26 inductor (used in Computer PSU), and HF decoupling caps. That should filter the noise from SMPS. You also have to separate the amp and SMPS at least 1ft away from each other, to reduce interference.
 
D

Deleted member 148505

Upcoming modules...

2 pairs Class D with sot23 totem pole buffers.

30A Speaker Protect with Fan Control.
 

Attachments

  • JLAmp1200D (3 x 6in).PNG
    JLAmp1200D (3 x 6in).PNG
    184.8 KB · Views: 3,161
  • Speaker Protect With Fan Ctrl (2.3 x 2.6in).PNG
    Speaker Protect With Fan Ctrl (2.3 x 2.6in).PNG
    103.9 KB · Views: 1,496
Status
Not open for further replies.