KG Balanced Dynahi build discussion thread

[spritzer]

I can take a stab at reducing the size a bit since those boards look to be gigantic. 

 

Local film PSU bypass caps as I used on the Dynalo Mk2.  They are on the amp boards to help make the PSU "faster". 

[Lil Knight]

Good stuff there

[spritzer]

Kevin sent me the board files so I'll see about reducing them.  Size is now about 230*160mm so difficult to fit sideways in a 19" chassis. 

[Lil Knight]

I don't mind longer boards, as long as the width is shorter.

[spritzer]

We are working on it....

[kevin gilmore]

this should be as small as we can get it.

please check for errors.

 

http://gilmore.chem.northwestern.edu/dynahibal8d.jpg

[jwzhan]

The amp section is giving me trouble.

I have only tested one amp and it was giving me pretty weird value (I don't remember the exact value now.) The LED off the negative rail is very bright and the negative rail measured something like -0.8V. The LED off the positive rail is very dim and flickers when the amp is powering off. The positive rail measured +30V.

I turned it on once for testing and haven't turned it on since. I figured that whatever is dropping the voltage is getting hot very fast and keep turning it on without knowing the reason isn't a good idea.

I did test the PSU output tho. It gave me the correct +30 and -30 when disconnected from the amp.

 

Like how I do with other things.... I made a LTSpice model for the amp.... I couldn't get the negative rail to drop like that no matter what I tried... weird... I guess I will have to turn it back on for more measurements now.

I'm praying that the other amp board won't give me any trouble.

[kevin gilmore]

pull the opamp

 

make sure that both 10k pots in parallel with the 500 ohm resistors are set at max resistance.

just noticed that one of them is going to be full ccw and the other is going to be full cw.

 

make sure the other 2 10k pots are set in the exact middle.

 

or start by not populating any of the 4 pots.

Edited June 28, 2013 by kevin gilmore

[jwzhan]

Thanks for the advice.

The amp board that I'm using is the off-board version.

I did have the pots parallel with the 500 ohm turned to max resistance.

I didn't set the pots at the input to the middle tho. I'll try that.

[kevin gilmore]

I've built a couple of those, they are known to work.

[kevin gilmore]

if you are building the older version with a lower voltage opamp and the regulators

for it, there are weird lock up issues with the regulators that cause the opamp

to do whacky things.

[ShinyFalcon]

I'm having a similar problem with jwzhan, the negative rail is ~0V and the positive rail is +30V with the Dynahi as load, negative LED slightly brighter than positive LED, both dim. On first power-up, no load, the 1N4007 diode on the negative rail was incorrectly soldered (- end on -30V, + end on gnd), so negative rail went to 0V. I swapped that diode against the silkscreen (+ end on -30V, - end on gnd) in agreement with the schematics on the first page, and voltage is +/-30V.

 

Also, if I install the 20k dual tracking resistor, negative rail goes to -40V. Maybe something blew in the process. I have another PSU that wasn't turned on before the diode swap, but waiting for some 10V references to come in.

 

I'll assume that the outermost trimpots are the ones in parallel with the 500 ohm resistors. Both of these reach max resistance of 1.3kohm (measuring outer pins) when both turned completely counterclockwise, and decreases when trimpot is turned clockwise. I can't measure the resistance of the middle trimpots, it reads a constant 0.8 kohm, so I just set them middle of the range. PCB revision is 0.2 for the PSU, and 1.0 for the Dynahi.

[kevin gilmore]

if its the original version of the power supply board, take out the current limit transistors on the

power supply board for both the positive and negative supplies.

[ShinyFalcon]

The board is the revision with the gate resistors. Removing the current limit transistors helped. The transistors did check out OK though, so I don't know. Shortly after installing the opamp I had to remove the positive rail transistors as well.

 

Anyways, amp is making music and sounds great!

 

Can unbalanced headphones be made to work on fully balanced amplifiers? I get lots of noise when I try on the balanced Dynahi, which I assume is due to the servo.

[kevin gilmore]

are you using a se input or a balanced input?

 

last time I checked, it was supposed to work all 4 ways.

[ShinyFalcon]

All the inputs are balanced. I left the inputs floating when I tested, potentiometer was already installed, but I haven't tested since then. I used O+ and GND, but all combinations of GND from star to earth ground didn't change the noise much, and sometimes louder on one side depending if headphone return floats or not. When I tested unbalanced phones the transformer was chirping occasionally, so maybe some instability.

[kevin gilmore]

it the source is true balanced with no ground reference then unbalanced output is not going to work with a high impedance pot

as that is the only source of ground reference. This amplifier is a true differential amplifier in this mode.

Edited July 8, 2013 by kevin gilmore

[holland]

For the onboard version, what is the bias limit w/ 2.5" heatsinks?  What is the transformer req, 60vct @ xxx?

 

Edit:  Well, assuming 160mA bias max, 200VA should cover it with headroom.  100VA if around 90mA.

Edited August 8, 2013 by holland

[Victor Chew]

Just got my offboard dynahi to work.

 

Had some initial teething problems with the PSU. It's working nicely now. To get it to work properly, I had remove the dual track and 6 limiters. Had to also repair the cut track under the board by jumpering from the 24v zener to the 2k resistor.

 

To get the amp DC voltage down I adjusted one of the outside 10k CCW and the other CW. With 7812, 7912 and opa27gp was able to get DC offset between 20mA to 40mA.

 

No music yet as input, output, ground wires and hardware yet to be sorted out. I will post pictures when more is done.

 

For me, this is more work and more challenging than either of my KGSShv builds.

 

Also, it runs quite hot as well. Amp side is about 62C and PSU about 52C at the heatsinks. Much hotter at the L brackets. Hoping the temperature will be lower after feet are installed and drilling of ventilation holes.

 

What kind of ground configurations are you guys using?

Edited September 29, 2013 by Victor Chew

[RudeWolf]

DC offset in mA? You should be trying to dial in the quiescent currents of the transistors to the correct values, wait for temps to settle and then hunt for zero mV DC on output. Then plug in the opamp and it should take care of the transients until normal working temps are achieved. If you are using soldered in opamps then be sure to measure DC at the opamp output pin rather than the output of the amp.

[Victor Chew]

Typo. Its mV DC.

[RudeWolf]

Is that at opamp outputs or amp? Ideally you should try to get zero mV on servo opamp outputs.

[Victor Chew]

For amp I made rough adjustments to about 200mV DC before slotting in the opamp. After which I got about 20 to 40 mV DC. Will be removing the regulators and using opa445 later.

[kevin gilmore]

some of the opamps have horrible offset specs themselves. as long as pin 6 of the opamp is in the range of +/-6v it is

doing its job.

[RodrigoRocha.RJ]

Does DynaFET will win a onbord version?