JLH Headphone Amp

Two versions now built ..

Hi guys

Busy week and have done LOTS of work this week on the boards, I have now build and tested two versions ..
  1. JHL Chinese Version
  2. JLH MkI Version

Lots of testing done with some very disappointing results, I have a very busy weekend of writing up todo ..
More soon

Miles

JLH MkI version from here
 

Attachments

  • JHL and JLH MkI.jpg
    JHL and JLH MkI.jpg
    361.5 KB · Views: 2,440
Last edited:
Testing the power supply on the PCB

Back on post #129 the board was assembled to the power supply stage and required testing.

The stage requires a good quality DVM, but a passive meter can be used at this stage.

Picture List
  1. Transformer Calculations
  2. Voltage Regulator pinouts

AC Test Points Assuming 12Volt Tx
  1. Set the meter to AC V
  2. AC (Left screw terminal) to GND= 13VAC
  3. GND to AC (Right screw terminal) = 13VAC
  4. AC (Left screw terminal) to AC (Right screw terminal) = 26VAC

DC Test Points
  1. Set the meter for DC V
  2. GND to 1N4007 (7912 Side of the PCB) bottom legs = -17VDC
  3. GND to 1N4007 (7812 Side of the PCB) bottom legs = 17VDC

Using the attached datasheet for the correct pintout, which is read as if the device number is facing you.
The device is highlighted in yellow TO-220 style case. Note the subtle difference in pinout!
Always keep the NEGATIVE or BLACK lead of the meter to GND therefore using the Positive or RED lead for the measurements.

7812 Test Points
  1. Set the meter for DC V
  2. Left Leg = 17VDC
  3. Centre Leg = 0VDC
  4. Right Leg = 12VDC

7912 Test Points
  1. Set the meter for DC V
  2. Left Leg = 0VDC
  3. Centre Leg = 17VDC
  4. Right Leg = 12VDC

Tips
  1. Don't forget to change the DVM from AC to DC for the second set of measurements.
  2. Take the time to satisfy yourself that the values are correct

To Done List
  1. Will Alps Blue Velvet fit the PCB ? - It will not physically fit on the PCB, it is just too big ..
  2. Is an output protection module required ?
    [*]What is a sensible transformer specification ? - 12 - 0 - 12 VAC at 25VA (Bigger VA number rather than smaller) See below
  3. Draw a logical and simple schematic

Next time, we finish the build of the PCB and start making noise ..
Hope this is helpful
Miles


DVM = Digital Volt Meter
Tx = Transformer
 

Attachments

  • JLH HA SbS - TX Spec.jpg
    JLH HA SbS - TX Spec.jpg
    216.8 KB · Views: 2,319
  • JLH HA SBS Build - 20b Pin out VReg.jpg
    JLH HA SBS Build - 20b Pin out VReg.jpg
    134.8 KB · Views: 2,148
Last edited:
Updated To Done List

To Done List

  1. Is an output protection module required ?
  2. Draw a logical and simple schematic
  3. Are larger heatsinks required ?
  4. Reduced the gain of the amplifier for a 2Vrms input rather than the older 0.707Vrms standard?

Have I forgotten anything else ?
More in the morning ...
Miles
 
Final Build Stages

During the final stages of assembly it is very important these components are inserted in the correct orientation ! Transistors and electrolytic capacitors suffer catastrophic failure in power the wrong way around ..

Picture List
  1. Input Connector and input transistor 2N2222 2 off
  2. Transistors - 2N2907 4 off
  3. Capacitors 2200uF - Post Regulator Smoothing 4 off
  4. Output Connectors
  5. Output Transistors TIP41 mounted on their heatsinks
  6. Output Transistors assemblies mounted on the PCB
  7. Potentiometer

Comments
  1. The 2N2222 and the 2N2907 are not interchangeable and must go in their correct positions and orientation. Looking from the top the device is in the shape of a D. On the PCB there is also a D shape and these flat sides allow the device to be inserted correctly.
  2. There is a slight misalignment between pins of the output transistor and the heatsink alignment holes on the PCB. The TIP41 pins are not quite straight when all is complete.
  3. The 2200uF capacitors are marked on one side with a minus sign and a shorter leg. The PCB is marked with a smile and this is the negative leg. Failure to insert any electrolytic capacitor with the correct polarity can lead to CATASTROPHIC consequences.

Tips
  1. Inserting the 2N2222 and the 2N2907 can be quite tricky and the legs will break if flexed tooo much.. I would suggest that the first and third legs are bent slightly out and the device then inserted with the bottom 5 to 7mm from the board.
  2. Can I suggest that the output transistors are mounted onto the heatsinks with the screws LOOSE... Insert on to the PCB and solder the middle pin of the device first, then check to see if the assembly is mounted square and flat to the PCB. Now solder the other legs of the transistor and the heatsink. Finally tighten the screw on the heatsink.
  3. The large power supply smoothing capacitors are another component which it can be easy to leave a gap between to the PCB. Again solder one leg and check that the component is tight to the PCB retouch the joint if necessary and then solder the other leg.

Now we are ready to power the board and finally test ...
Miles :D
 

Attachments

  • JLH HA SBS Build - 23 2N2222 and input.jpg
    JLH HA SBS Build - 23 2N2222 and input.jpg
    183.2 KB · Views: 2,062
  • JLH HA SBS Build - 24 2N2907.jpg
    JLH HA SBS Build - 24 2N2907.jpg
    182.5 KB · Views: 1,925
  • JLH HA SBS Build - 25 2200uF Post VReg.jpg
    JLH HA SBS Build - 25 2200uF Post VReg.jpg
    152.4 KB · Views: 381
  • JLH HA SBS Build - 26 Output Connectors.jpg
    JLH HA SBS Build - 26 Output Connectors.jpg
    163.2 KB · Views: 345
  • JLH HA SBS Build - 28 Pot.jpg
    JLH HA SBS Build - 28 Pot.jpg
    166.1 KB · Views: 400
  • JLH HA SBS Build - 27b Output Tx's PCB.jpg
    JLH HA SBS Build - 27b Output Tx's PCB.jpg
    168.8 KB · Views: 375
  • JLH HA SBS Build - 27a Output Tx's Mounted.jpg
    JLH HA SBS Build - 27a Output Tx's Mounted.jpg
    212 KB · Views: 353
Last edited:
Testing the completed board

I have tried to make this testing phase simple and painless, with all of the testing done with a DVM.
Some measurements are taken With Reference to Ground WRG Using the black test probe connected to the GND terminal of the AC input connector.

Before the board is powered a load should be applied across the output terminals, in my case 100ohm resistor at 1 watt. Please don't use your best headphones at this stage, if something is wrong this could be a VERY expensive mistake.

I have assumed that the board is on the bench with the AC connector at the top as shown in the picture.

Picture List
  1. Test rig Two Avo 8's measuring the AC current into the board.
  2. AC Input connector
  3. Pinout of the voltage regulators
  4. 5.1 Ohm resistors
  5. Zener Diodes
  6. Finally the Output Connector

AC Test Points
  1. With the meter set to AC volts
  2. AC input connector WRG approx 13V at each AC input
  3. AC input connector approx 26V at between each AC input
  4. Voltage Regulator output pins WRG - a few millivolts
  5. Output Terminal connector of each channel WRG less then 5mV

DC Test Points
  1. With the meter now set to DC Volts
  2. Right side 1N4001 bottom pin WRG 17V DC
  3. Left side 1N4001 bottom pin WRG -17V DC
  4. Output pin 7912 WRG -12V DC
  5. Output pin 7812 WRG 12V DC
  6. Across the 5.1ohm resistor 0.6V DC (Red test probe top and black test probe bottom)
  7. Across the zener diode 3V DC (Black test probe top and red test probe bottom)

If all of the above measurements are correct, the heatsinks of the devices should start to warm up.. After about 5 mins a finger placed on top will feel heat. Both voltage regulators have over temperature, short circuit and over current shut down if anything is seriously wrong.

If all the above conditions are met the final stage of setup can be completed below.

Setup
The variable resistor is a 10 turn version which removes the DC offset at the output. With the DVM connected to the output connector with variable resistor should be turned until the offset is removed..

Comments
Please note : A limitation of this design is that the DC offset will vary over time and due to heat. 50mV = 0.05V is quite acceptable

At this point is VERY tempting to plug your best headphones in and kick back with your favourite CD .. Please don't .. Can I suggest that the board is now soak tested for at least 6 to 10 HOURS to allow any suspect components to fail .. If a semiconductor or electrolytic capacitor is to fail it will happen in the early period or end of a product life cycle.

Check the DC offset value once again and now FINALLY kick back and listen to your hard work ...

Next I will look at some AC testing of the board..
I hope this has been helpful
Miles

DVM = Digital Voltmeter
 

Attachments

  • Testing - 1 Test Rig.jpg
    Testing - 1 Test Rig.jpg
    96.6 KB · Views: 730
  • JLH HA SBS Build - 19 AC in.jpg
    JLH HA SBS Build - 19 AC in.jpg
    188.4 KB · Views: 699
  • JLH HA SBS Build - 20b Pin out VReg.jpg
    JLH HA SBS Build - 20b Pin out VReg.jpg
    134.8 KB · Views: 438
  • JLH HA SBS Build - 12 5R1.jpg
    JLH HA SBS Build - 12 5R1.jpg
    189.8 KB · Views: 401
  • JLH HA SBS Build - 04 Zener Diodes.jpg
    JLH HA SBS Build - 04 Zener Diodes.jpg
    205.1 KB · Views: 365
  • JLH HA SBS Build - 26 Output Connectors.jpg
    JLH HA SBS Build - 26 Output Connectors.jpg
    163.2 KB · Views: 395
AC Testing

Below are a series of waveforms to test the design under ideal conditions by applying an impulse to the input, in this case a square wave and the output is displayed on an oscilloscope. The oscilloscope has two channels CH1 and CH2 which have the same functionality. The X10 (Times 10 or multiplied by 10) versions show an expanded X axis of the graph to quickly show the effective rise time of the amplifier.

Picture list
  1. 100Hz Square wave
  2. 100Hz Square wave X10
  3. 1kHz Square wave
  4. 1kHz Square wave X10
  5. 10kHz Square wave
  6. 10kHz Square wave X10

In the following tests CH1 (Top of the screen) displays the input waveform from the signal generator and CH2 (Bottom of the screen) displays the output of the amplifier. In a perfect world both signal should look the same and the rise time of each signal should match.

Testing the board
A square wave of the following frequencies is applied to the board and the output recorded.
  1. 100Hz
  2. 1HKz
  3. 10kHz

On the right hand side of each picture there are a number of measurements taken by the scope which are helpful to understand what is happening..
  1. The first is the frequency in Hz of the input signal
  2. The second is the rise time of the orignal signal on Channel 1 CH1
  3. The third is the rise time of the output of the amplifier on Channel 2 CH2


Comments
There are 6 pictures, 3 sets of two for the following frequencies
  1. 100Hz Picture 1 and 2
  2. 1HKz Picture 3 and 4
  3. 10kHz Picture 5 and 6

Conclusion

The sharp increase in rise time and the X10 pictures clearly show the reactance of the amplifier rapidly increasing as the frequency increases, in effect reducing the high frequency performance.

So this Chinese version of the JLH design will sound dull and a bit lifeless ..

It must be strongly emphasised that the original design was capable of much better performance than this implementation. So the next job is now convert this board to the original JLH shown in circuit 1 here.


Please note : My waveform generator was made a technical college many years ago and had the latest 8038 device at the centre of the design.. Many years have passed and it can be seen that the square output at low frequencies is not very good at all, so a new version is required..

Picture list
  1. 100Hz Square wave
  2. 100Hz Square wave X10
  3. 1kHz Square wave
  4. 1kHz Square wave X10
  5. 10kHz Square wave
  6. 10kHz Square wave X10
 

Attachments

  • Testing - 3a 1kHz.jpg
    Testing - 3a 1kHz.jpg
    100.4 KB · Views: 259
  • Testing - 2b 100Hz x10.jpg
    Testing - 2b 100Hz x10.jpg
    100 KB · Views: 289
  • Testing - 2a 100Hz.jpg
    Testing - 2a 100Hz.jpg
    100.7 KB · Views: 446
  • Testing - 3b 1kHz x10.jpg
    Testing - 3b 1kHz x10.jpg
    96.2 KB · Views: 243
  • Testing - 4a 10kHz.jpg
    Testing - 4a 10kHz.jpg
    102.4 KB · Views: 261
  • Testing - 4b 10kHz x10.jpg
    Testing - 4b 10kHz x10.jpg
    98.5 KB · Views: 257
JHL In Summary

PCB Issues
  1. C23, C24 : Are shown on both the schematic and the PCB with incorrect polarity
  2. The left and right swap sides on the PCB - the input connector is logically labelled L - G - R, however the LEFT channel is on the RIGHT side of the PCB and the RIGHT channel is on the LEFT side of the PCB!
  3. The ground track for both outputs is a VERY poor layout
Component Issues
  1. C9, C10 : 33nF However the supplied value is 0.1uf
  2. C13, C14 : 1uF Are supplied as 0.47uF
  3. C15, C16 : 330pF Is a ceramic capacitor which will perform poorly in this function
  4. C1, C2 - C23, C24 : Would benefit from having high quality components
By design the original JLH version was inherently stable at RF frequency a number of components are not required
C5, C6 - C19, C21 - C17, C18

The zener diode stated on the schematic diagram 2V7 = 2.7V The supplied value is 3V0 = 3V

The zero ohms links supplied are of poor quality and have a value of 0.4 to 1 ohm

In Summary
C23, C24 Require inserting with the correct polarity

There is also another layout issue around the input of each channel with regard to space C13, C14 - C15, C16 It seems a small ceramic capacitor of 330pF has been used to allow a fit next to the MUCH larger 0.47uf which by design should be a 1uF again would be larger still.

Some of the above issues with the PCB cannot be resolved and have to be put down to experience.

In an attempt to make the best of the current PCB and components available another two versions will be built
  1. JLH Original using the circuit 1 from here - Done and Tested
  2. JHL Simplified with correct component values - This month July 2011

To Done List
  1. JHL Simplified with correct component values - This month
  2. Is an output protection module required ? - This month
  3. Draw a logical and simple schematic - When the circuit is finalised
  4. Reduced the gain of the amplifier for a 2Vrms input rather than the older 0.707Vrms standard
  5. Are larger heatsinks required ? - Due to the reduction in gain this may not be an issue
 

Attachments

  • JHL HA SbS PCB - Combined Layers.jpg
    JHL HA SbS PCB - Combined Layers.jpg
    465.8 KB · Views: 726
  • JHL Headphone Amp MKI Schematic.pdf
    48.5 KB · Views: 766
Cleaning up the desktop

Hello all

Just a couple of points that I keep meaning to add before another installment of the JHL build on
  1. Alps Blue Velvet Potentiometers
  2. Heatsinks
  3. and finally transformers

Upgrading the Volume Potentiometer mounted on the PCB
There a number of areas on the PCB where there are size constraints and around the Volume Pot there is very little space. Reviewing the attached datasheet shows that this element would need to be an "off board" modification.

Increasing the sizes of the heatsinks
The form factor of the heatsinks used would allow an increase in size if required. The attached drawing shows that the width and depth are fixed and the height is changed for the application.

Transformer Selection
The choice of 25 or 50VA transformer was based apon the regulation of the size of transformer, broadly speaking the larger the transformer the better the regulation. Further information can be found on the attached datasheet.

All of the above information has been taken from Rapid Electronics here in the UK. Over the past 20 years without exception I have always found them profession, helpful and a great company to deal with. (also their catalogue and website are much simplifier and smaller than others we can mention)

With all the future modifications I have tried to use easy to find worldwide sourced components from the big companies.

Right onto the new work and modifications to the JHL PCB
Miles
 

Attachments

  • Alps Blue Velvet.pdf
    87.6 KB · Views: 282
  • To-220 Heatsink.gif
    To-220 Heatsink.gif
    9.6 KB · Views: 2,396
  • Vigortronix Transformer from Rapid Electronics SMALL.pdf
    172.7 KB · Views: 248
Last edited:
From JHL to JLH

Picture List
  1. Incorrect capacitor polarity issue
  2. JHL to JLH - The PCB's side by side
  3. Component locations for change
  4. 100Hz Square wave
  5. 100Hz Square wave X10
  6. 1kHz Square wave
  7. 1kHz Square wave X10
  8. 10kHz Square wave
  9. 10kHz Square wave X10

Comments
The performance of the JHL kit was not as expected or designed by JLH and having serious reservation with regard to the PCB layout proof was required that the project should continue. The work done for this post was to improve the amplifier performance at little or no cost and to check the PCB for function.

Limitations in testing
It must be made very clear that the test regime for the amplifier is currently very basic with a number of omissions, these are
  1. Poor input test signal quality at low frequencies
  2. The full bandwidth has not been tested
  3. The input signal has been limited to a small value of 200mV peak to peak
  4. Testing at 20, 200, 2k and 20kHz would be more useful and the design improves
  5. Both sine and square testing at the above frequencies
Steps are now being taken to improve the testing setup by serval orders of magnitude.

Stage 1
These simple changes will improve the performance of the board considerably.
  1. C23, C24 Require inserting with the correct polarity. The negative electrolytic capacitor terminals should face the centre of the board. If the board has been powered up can I suggest this component is replaced.
  2. Remove the 8 (0.1uf) capacitors shown in photo 3

Stage 2
With a small number of replacement components the design takes another step forward. Using picture 3 as our guide
  1. Implement Stage 1
  2. Change 2k2 for 3k3 resistors (2 off)
  3. Remove the two 10k variable resistors and link as required
  4. Replace 4k7 with 10k (2 off)
  5. Replace 220k with 47k (2 off)
  6. Replace the Zener Diode with 3 x 1N4148 in series (6 off) - Note the polarity is opposite to that of the zener.

Please note that there is a larger offset voltage which cannot be removed with this method in the order of ±100mV.

Impulse Testing
Below are a series of waveforms to test the design under ideal conditions by applying an impulse to the input, in this case a square wave and the output is displayed on an oscilloscope. The oscilloscope has two channels CH1 and CH2 which have the same functionality. The X10 (Times 10 or multiplied by 10) versions show an expanded X axis of the graph to quickly show the effective rise time of the amplifier.

In the following tests CH1 (Top of the screen) displays the input waveform from the signal generator and CH2 (Bottom of the screen) displays the output of the amplifier. In a perfect world both signals should look the same and the rise time of each signal should match.

There are 6 pictures, 3 sets of two for the following frequencies
  1. 100Hz Picture 4 and 5
  2. 1HKz Picture 6 and 7
  3. 10kHz Picture 8 and 9

On the right hand side of each picture there are a number of measurements taken by the scope which are helpful to understand what is happening..
  1. The first is the frequency in Hz of the input signal
  2. The second is the rise time of the orignal signal on Channel 1 CH1
  3. The third is the rise time of the output of the amplifier on Channel 2 CH2

In Conclusion
These small changes have made a considerable improvement to the JHL kit with respect to the rise time of the headphone amplifier and therefore bandwidth. So in real terms there should be more detail and life in the top end..

Next Step
To fully implement the original JLH design with the correct capacitor values and change the output transistors to the original 2N492x. These components are currently on order.. In real terms these are small changes but personally I am very interested to see the difference they will make.

Picture List
  1. Incorrect capacitor polarity issue
  2. JHL to JLH - The PCB's side by side
  3. Component locations for change
  4. 100Hz Square wave
  5. 100Hz Square wave X10
  6. 1kHz Square wave
  7. 1kHz Square wave X10
  8. 10kHz Square wave
  9. 10kHz Square wave X10
 

Attachments

  • JHL HA SbS - JHL and JLH MkI.jpg
    JHL HA SbS - JHL and JLH MkI.jpg
    361.5 KB · Views: 2,386
  • JHL HA SbS PCB - Capacitor Issue.jpg
    JHL HA SbS PCB - Capacitor Issue.jpg
    465.8 KB · Views: 2,462
  • JHL HA SbS - JHL to JLH changes small.jpg
    JHL HA SbS - JHL to JLH changes small.jpg
    698.5 KB · Views: 2,353
  • JHL HA SbS - Testing JHL_JLH 100Hz.jpg
    JHL HA SbS - Testing JHL_JLH 100Hz.jpg
    221.5 KB · Views: 2,279
  • JHL HA SbS - Testing JHL_JLH 100Hz X10.jpg
    JHL HA SbS - Testing JHL_JLH 100Hz X10.jpg
    216.6 KB · Views: 317
  • JHL HA SbS - Testing JHL_JLH 1kHz.jpg
    JHL HA SbS - Testing JHL_JLH 1kHz.jpg
    214.1 KB · Views: 279
  • JHL HA SbS - Testing JHL_JLH 1kHz X10.jpg
    JHL HA SbS - Testing JHL_JLH 1kHz X10.jpg
    212.1 KB · Views: 247
  • JHL HA SbS - Testing JHL_JLH 10kHz.jpg
    JHL HA SbS - Testing JHL_JLH 10kHz.jpg
    203.8 KB · Views: 262
  • JHL HA SbS - Testing JHL_JLH 10kHz X10.jpg
    JHL HA SbS - Testing JHL_JLH 10kHz X10.jpg
    191.9 KB · Views: 298
After modding to stage 1, i don't hear the différence but after changing R to reduce the gain and well ordered the offset, Yes it's seems to me that the sound is better, more bass, voice are better defined. Thanks for you help.
- Transformer: i'm going to by a 50VA Rcore 2x 12v because mine is too hot
-change the n4007 by BYW 98-200, some blackgate capacitors and two muse on the output
-For the ALPS, it must be deported on the panel, because it's too small for implant it on the PCB
Regards
 
Want to thank Miles for sharing his journey with such thoughtfulness, detail and clarity. I'm looking for a good headphone amp to build and the JLH design may be it, or even the JHL version, so I'm looking forward to see if Miles can redeem that board :)

May I throw in my 2c here and suggest instead of an Alps pot that the Lightspeed attenuator be considered? I've found LDRs to be so much more transparent sounding than mechanical pots, stepped attenuators or TVCs.
 
Hello all

This week I have completed another two configurations of the JHL->JLH board with testing and I'm pleased to report that we have progress. It should be clearly understood the original design is now antique in the terms of electronics and audio design but has stood the test of time very well. However modern sources eg CD players are designed for a 2Volt peak to peak output (1.4VRMS) with programme content averaging 300-500mV, JLH designed this headphone amplifier with a 0.707VRMS or 1Volt peak to peak 100-200mV.

There is a mistake on post #147
To Done List - Point 4. Reduced the gain of the amplifier for a 2Vrms input rather than the older 0.707Vrms standard
Should read - Reduced the gain of the amplifier for a 2Vpeak to peak input rather than the older 0.707Vrms standard

To Done List
  1. JHL Simplified with correct component values - Done
  2. Reduced the gain of the amplifier for a 2Vpeak to peak input rather than the older 0.707Vrms standard
  3. Investigate JLH original 2N492x output transistor
  4. Investigate the ability of the power supply to perform and will this require an output protection module required ? - This month
  5. Are larger heatsinks required ? - Due to the reduction in gain this may not be an issue
  6. Draw a logical and simple schematic - When the circuit is finalised
  7. Investigate Lightspeed attenuator

fj12centauro - I look forward to your opinion when all is complete. Can I suggest waiting before you buy a new transformer as there is a pending question on the power supply.

Bernie7 - Thank you for your kind words. I have added the attenuator to the growing list.

jambul - Interesting question! By performance do you mean bandwidth/frequency response? The new testing seems to indicate that the current design is good, but I want to investigate why JLH chose the 2N4922 transistor. He did everything for a reason and I need to understand this decision, however I think it is due to current gain of the final stage. More on this very soon.

Over the weekend I will write up the testing completed this week
  1. Tested Original JLH the correct values of C and R - JLH C&R
  2. Tested JHL Stage 1 with corrected C&R - JHL Mod C&R

Soldering irons at the ready please.

Miles
 
Update

Hello all

Just to follow up from yesterday on a couple of points ..

fj12centauro - Replacing the 1N4007 is interesting. After much research the BYW98-200 was made by ST but seems to have been obsoleted, the datasheet is easily found but the device is missing from the ST website.

Bernie7 - The attenuator uses a Silonex NSL-32SR2 Optocouplers. This seems to be a VERY cheap and interesting way to control the volume, the devices are inexpensive and available from Farnel ..

Testing
The testing rig used is really starting to show it's age and limitations now the design is starting to develop, as mentioned previously the signal generator was made at college and the scope has an issue with a DC offset on channel 2. Also it is now becoming important to measure the bandwidth and THD figure of the headphone amplifier as these were the tenants on which the design was conceived.

Dummy Load
All the testing completed so far has used some pretty horrid (Technical term) wirewound 100ohm 1 watt resistors similar to that used by JLH in the original articles. This part of the testing needs to be improved across the various values in the headphone market today
  • 10 ohm
  • 68 ohms
  • 330 ohms
  • 680 ohms
  • 10 ohm + 0.22uf
  • 68 ohms + 0.22uf
  • 330 ohms + 0.22uf
  • 680 ohms + 0.22uf
The resistors used will be the standard 6watt wirewound variant commonly found in distribution. JLH also tested with a resistive and capacitor load of 0.22uF this will give a worse case scenario rather than pretty test figures.

The components that are being selected for the design and testing are hopefully available worldwide so that the work can be replicated in the future regardless of location.

That's all for now
Miles
 
Last edited:
Bernie7 - The attenuator uses a Silonex NSL-32SR2 Optocouplers. This seems to be a VERY cheap and interesting way to control the volume, the devices are inexpensive and available from Farnel ..

Miles

Yes it is. You may be surprised as I was at how much mech wiper pots add to what we commonly attribute to the dry, grainy and etchy 'transistor sound'.
 
Last edited:
Finally...

After a bit of testing, I've settled on increasing the gain resistor, R5 on the JLH schema, to 2K. This reduces the overall gain. The input capacitor has been increased to it's proper value of 1uf. C2 has been replaced with a low ESR type, and C4 has been increased to 220uf low ESR (both numbered by JLH).
Using a 15-0-15 AC 30VA transformer the regulator heatsinks are at about 60C (ambient is 23C), the TIP41s are at about 50C. Offsets are varying between +/- 5mV.
The board is on final soak test before being boxed and classed as 'done'. It sounds very good, and a bargain for the cost.
Next on the chopping board is the Panda headphone amp. Damn these cheap kits, I've got the 'bug' I thought I'd cured years ago. I've a nice workshop with decent equipment now though! (Question: Should all this equipment and components etc. be added on to the cost of the kits? If so, for me, they're crap value for money :D)
 
Last edited: