RJM Audio Sapphire Desktop Headphone Amplifier

Fuses blowing even without the amp boards connected? Slow blow types? If so I'd be worries that the fuses are doing their job and there's a short circuit in the power supply...

i am using a DMM to measure the output voltage at the output plug and after a little while the fuse would blow

the transformers draw 1100ma unloaded, two of them would probably blow a 500ma fuse :)

I might go up to a 2.5a fuse, sounds about right?

-joe
 
Fuses blowing even without the amp boards connected? Slow blow types? If so I'd be worries that the fuses are doing their job and there's a short circuit in the power supply...

I am using the 62032 transformers like your prototype.. what size fuse did you use?

UPDATE on previous.. my mistake.. it's not 1100ma unloaded.. it's 5.0ma.. so i am thinking is that when I use the DMM to test the voltage, it puts a load so it draws 1041ma (according to the spec sheet) over the 500ma fuse limit

-joe
 
Last edited:

rjm

Member
Joined 2004
Paid Member
Looks great Joe.

How are you handling the connection between the board common and the chassis?

Normally this is made at the output jack where the out- from each board connect to the common GND pin, a third wire is connected from the GND pin to the chassis.

In any event, out- from each board should connect to the chassis somehow.
 

rjm

Member
Joined 2004
Paid Member
I trying to catalog the issues people have encountered when building the Sapphire amp, with the view to changing the documentation and guides to help prevent repeats.

problem: 500 mA fuses blow on inrush current
fix: 1 A slow-blow (T) fuses recommended

problem: damaged transistor Q3 (overheat on soldering) caused DC offset of 8 V or so, potential to kill 16 ohm headphones
fix: care on soldering, correct orientation and ID of transistors, check all voltages before connecting headphones

problem: voltage too high/low
fix: voltage is adjusted by R7, and set to be a compromise, a little low for 300 ohm phones, a little high for 16 ohm phones. It can be optimized for one extreme or the other but not both. The solution is to increase R14 to 33 ohms or higher, which also acts as a current limiter, protecting headphones from damage.

problem: jumper setting to put buffer in closed loop config doesn't seem to work.
fix: until some people actually try to get this to work, there's not much I can add to the discussion. It should be possible, but additional parts such as capacitance in parallel with the feedback resistor might be required. Some trial and error will be involved.
 

rjm

Member
Joined 2004
Paid Member
sorry, above should read instead,

problem: volume too high/low
fix: volume is adjusted by R7, and set to be a compromise, a little low for 300 ohm phones, a little high for 16 ohm phones. It can be optimized for one extreme or the other but not both. The solution is to increase R14 to 33 ohms or higher, which also acts as a current limiter, protecting headphones from damage.

Also: going to order more sapphire boards, kits available if there is demand for them. Let me know if interested.
 
I know that the pcb was design as a "drop in replacement" but have you given thought to combining the L+R to one board? the wiring would be must simpler

Also if there is room, if we can use mur860 or something so that we can have the transformer hooked up directly without using bridges?

-joe
 

rjm

Member
Joined 2004
Paid Member
Switching op amps

After using the OPA124 since the Sapphire was first built, I finally got around to swapping in a pair of OPA27 instead.

The circuit should be pretty much agnostic with respect to different "audio" (c. 10MHz) op amps. Bandwidth, noise floor, and stability should be unchanged: since the op amp does not see the output load (headphones).

The difference between the two op amps is the OPA134 is a FET input circuit optimized for low distortion while the OPA27 is a bipolar input circuit optimized primarily for low noise.

Given the relatively high input impedance presented by the volume control, one might imagine that the FET-based insput stage would work out better.

I was seriously wondering if I'd be able to hear any difference at all, but in the event it was quite easy to pick out.

The OPA134 does, I think, sound better. The OPA27 is hardly bad, some people might actually prefer it, but while slightly softer it is also more congested and constrained to a more local, denser soundstage. With the OPA134 the music jumps out and around, more blackness, more three-dimensional. The OPA134 might be a bit over-cooked, I'm still considering this.

Side note: with the OPA27, the Sapphire sounds almost exactly the same as the 47 labs 0247, also a bipolar design.

Anyhow, the op amp in the Sapphire is socketed, playing around with different types is a cheap, safe, and easy tweak.