O2 amp CRC, diode, cap, and heatsink mods

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
O2 amp - DC power jack mod corrections and 12Vdc mod

Here are a couple of corrections and additions on the DC power jack charging-only mod above, plus how to use it with 12Vdc instead of 24Vdc using a DC-DC converter chip.

* I have the new diode D8 in the wrong place on the schematic I posted above. The corrected schematic is below. The new diode D8 goes in series from the center pin of the new DC jack to pin 1 of the O2 power switch.

* I'm changing the part on the new diode D8 from a 1N5818 Schottky to a 1N5919:

Mouser 1N5819 Schottky Diode

The 1N5819 is rated at 40V vs. 30V for the 1N5818 used in the O2. That buys an extra 10Vdc of safety margin in the unlikely case a certain 24Vdc adapter should exceed 30V when unloaded.

* I want to make it a bit clearer that the only "new" parts here are the D8 diode and the J4 DC power jack. Everything else in the schematic is just a snippet of the O2 schematic to show how they wire up.

* The J4 DC jack needs to be a fully insulated type, that does not allow either power jack terminal to come in contact with the case, when the jack is panel mounted. The case of the O2 will be at ground potential while the two DC jack leads will be at rail voltages. The jack also needs to be rated at 30V or higher. This 2.1mm jack from Mouser would do the job:

Mouser 2.1mm 100V max, fully insulated, DC jack

DC jack data sheet

* I've lowered the minimum current requirement of the 24Vdc adapter to 50mA. The maximum charging current with both batteries discharged to around 7Vdc and with a "24Vdc" adaptor running at 30Vdc would be:

[30Vdc - (2 * 7Vdc)] / (2 * 220R) = 36mA

Then decreasing to trickle rates as the batteries charge.

* This DC jack modification for charging-only use would probably have the most use in situations where the DC is already available, like a car, given that the 24Vdc 50 mA adaptor would be about the same size as the small WAU12-200 AC adaptor in the O2 BOM, as noted above. Not much to be gained in adapter size if the goal is to plug something into the wall outlet.

This O2 amp DC charging mod could actually be used with 12Vdc from a car if a 12VDC-to-24Vdc single rail converter were used, as shown in the second schematic below. A dc-dc converter like this one would do the job:

DigiKey 12V to 24V DC-DC converter, 2W 80mA

DC-DC converter data sheet

This particular converter has a fairly wide input voltage range of 9Vdc to 18Vdc, to handle the system voltage fluctuations in a car. The schematic shows the input and output capacitors needed. The 1.4k output resistor is to maintain the minimum 8mA output current load required by the converter, even when the converter is switched out of the circuit by the power switch.
 

Attachments

  • O2 add DC port 2.png
    O2 add DC port 2.png
    35.2 KB · Views: 1,061
  • O2 add DC port 3.png
    O2 add DC port 3.png
    50 KB · Views: 1,058
Last edited:
The other way to do this is to float charge the pair of batteries with DC at the max rate for your batteries (generally 1/10th their rated mAH capacity--i.e. for 200 mAH batteries set the charge rate to 20 mA). You can use a regulated (switching) or unregulated 24 VDC wall transformer. The advantage:

The DC power supply will supply most of the amp's power requirements when the amp is operating. Using AGDR's method above the amp runs only from battery.

The disadvantage over using an AC wall transformer:

The O2 will still eventually run its batteries dead with enough use even with the DC power supply connected unless you set the charge rate high enough to cover its average power consumption (typically under 30 mA but can be higher if you like to listen loud with power hungry headphones).

Using a regulated supply you could make the supply more "self tapering" much like the O2's internal battery charging.
 
Thanks, RocketScientist! :) I was thinking about that method, but it didn't appear that NiMH has a float voltage, as per

How to charge Nickel Metal Hydride Batteries.

but lead-acid gel cells certainly do. An O2 could be mounted on a vehicle and fed by a couple of 12Vdc UPS batteries in series to replace BT1 and BT2, each fed with an (isolated) float charger.

Below is another way to do DC charging, using a 12Vdc DC-DC converter that outputs +/-12Vdc this time instead of +24Vdc. In this configuration a ground is produced that is connected to system ground. The +/-12Vdc outputs then go to the power switch pins 1 and 4 with the addition of a second protection diode, D9.

I've updated the schematic for the 12Vdc -> +24Vdc DC-DC converter case with some additions and better values around the converter. I've included the noise reduction capacitors and inductors here from the data sheet, although there is not much benefit there since the audio section of the O2 is switched out while the batteries are charging.

Both the these specified DC-DC converters are isolated. It would not matter if the internally-grounded metal case of the O2 touches the metal chasis of the car feeding the DC-DC converter. No short circuit occurs.
 

Attachments

  • O2 add DC port 3.png
    O2 add DC port 3.png
    31.7 KB · Views: 1,010
  • O2 add DC port 2_a.png
    O2 add DC port 2_a.png
    27.6 KB · Views: 992
Last edited:
What I'm suggesting is simply using a 24 VDC power supply to charge the batteries at roughly the same rate the onboard resistors (R1/R2) and 12 volt regulators/diodes do. The amp will charge when turned off, and only use a slight amount of battery power when on and playing at low volumes. It would only put significant drain on the batteries if you worked it hard into low impedance hungry headphones.

The battery life, plugged into such a 24 VDC supply, should easily last even a full 12 hour day and the remaining 12 hours at night should be enough to top up the batteries for the next day. So it's a bit of "wear and tear" on the batteries but it gets around the need for an AC wall transformer.

To get technical, I perhaps shouldn't use the term "float" in this case. It's probably more accurate to say "trickle charge". You could set up a regulated supply that would let the charge current taper to near zero when the NiMh batteries were fully charged but that's tricky and very battery specific. And it still needs to either have active current limiting to C(mAH)/10 or a resistor to accomplish the same thing. The reason is NiMh batteries, unlike lead acid, drop in voltage if they're overcharged. So without current limiting, they can start drawing significant charge current again after they taper off and reach full charge. That's how most smart chargers work. They detect the "peak" and shut off or drop to trickle charging once the voltage starts dropping.
 
O2 amp DC servo modification

Here is another O2 modification that is completely unnecessary, since the O2 already has a very low 3mV-or-so output offset voltage, but fun to mess with anyway. :) This modification adds a DC servo to each pair of output buffers on the O2 to zero out that small remaining offset voltage.

The source of the tiny 3mV DC offset in the O2 is the input bias current from the NJM4556 op amps going through the 40.2K input filter resistors, producing a 3mV offset that gets reflected to the output by the 1:1 buffer configuration.

To see this in action, as a fun test, substitute the two NJM4556 chips with two LME49720 (same as LM4562 and the higher voltage LME49860) chips and measure the DC offset, with no headphones attached. With LME49860 chips I measure 0.4mV (400 micovolts!) on one channel and actually 0.0mV on the other, meaning the offset voltage is below the least significant bit of the meter. This happens because the input offset voltage of the LME chips are 1/5 that of the NJM chips from the datasheet specs. Note the LME chips can't handle anywhere near the output current the NJM4556s can, so they would only be useful substitutes for 600R headphones, but this test just shows the effect of op amp input bias current on DC offset into and out of the O2 buffer stage.

The DC servo circuit is a simple integrator similar to what was used in the famous Dynahi headphone amplifier on Headwize.

DIY RESOURCES - Kevin Gilmore DYNAMIC HEADPHONE AMPLIFIERS and POWER SUPPLIES

The operation is simple. The output of the O2 channel 1 (past the 1R resistors) feeds the DC servo integrator. The 40.2K input filter/bias resistor on the O2 buffer pair for that channel is then returned to the resultant DC output of the integrator, rather than signal ground, to complete the DC servo feedback loop.

In the LT SPICE simulation below I've added current sources to the input of each buffer pair to forcibly inject a DC offset. I've added the DC servo circuit to channel 1, but not channel 2, then injected the same extra input offset current into both channels to compare the effects with the DC servo (Ch1) and without the DC servo (Ch2).

In the first plot a 10uA injected current results in a 200mV input offset voltage from the output of O2 channel 2. This is 70 times higher than the normal 3mV O2 DC offset, but I used the large value to make the effects visible on the plots. The red plot is the input signal, centered around zero of course. I've only shown the input signal for one channel since they are identical and would just sit right on top of each other. The blue plot is the output of Ch2 showing it shifted down by -200mV in response to the injected 10uA current. And the green plot is the output of Ch1 with the DC servo attached, showing the servo has corrected for the same 10uA current injection resulting in a perfectly centered (0mV offset) output waveform. The magenta plot is the DC servo output (DC level) from the integrator being fed back to the O2 buffer stage.

The second plot shows the results with an even more extreme 30uA of injected offset current, 3 times the about above. Once again the DC servo circuit on Ch1 has correctly neutralized the offset voltage, producing a centered waveform, vs. Ch2 without. Note the larger DC error signal (magenta plot) being produced by the DC servo integrator in response to the larger DC offset that needs correction.

The first circuit diagram is the O2 signal path showing the DC servo added to Ch1. The second diagram is just a closeup of the servo portion. I've only shown the one DC servo on the diagram to compare with the second channel without. A total of two servo circuits would be needed, of course, one for each channel. A dual op amp chip could be used for the DC servo amplifier, one section for each O2 output channel. I used a polar feedback capacitor in the integrator in the sims, but an actual circuit should have a non-polar cap here since the integrator output can swing either positive or negative as needed. 3 3.3uF 25v MLCC XLR ceramics in parallel would be a good way to go.

The first two plots (10uA and 30uA injected DC offset current) are both done at 2.5kHz. The last two are for 10uA offset current injection done at 20Hz and 25kHz to show the servo still works at the frequency extremes, since the integrator is frequency dependent, of course. The DC servo output level looks a little bumpy in the 20Hz plot, but remember this is a 200mV offset signal that is being corrected, just to make it visible. The actual O2 DC offset is around 3mV.
 

Attachments

  • O2 amp DC servo ILOAD=30uA on Ch1.png
    O2 amp DC servo ILOAD=30uA on Ch1.png
    27.1 KB · Views: 88
  • O2 amp DC servo ILOAD=10uA on Ch1.png
    O2 amp DC servo ILOAD=10uA on Ch1.png
    27.8 KB · Views: 140
  • O2 amp DC Servo closeup on Ch1.png
    O2 amp DC Servo closeup on Ch1.png
    12.9 KB · Views: 168
  • O2 amp signal path with DC Servo on Ch1.png
    O2 amp signal path with DC Servo on Ch1.png
    43.2 KB · Views: 963
  • O2 amp DC servo ILOAD=10uA on Ch1 20Hz.png
    O2 amp DC servo ILOAD=10uA on Ch1 20Hz.png
    27 KB · Views: 77
  • O2 amp DC servo ILOAD=10uA on Ch1 25khz.png
    O2 amp DC servo ILOAD=10uA on Ch1 25khz.png
    27.6 KB · Views: 78
Last edited:
How do you calculate the injection?

Good question! All the injection current gets forced through that 40.2K resistor. So for 10uA that gives (10uA)(40.2K) = 400mV, which brings up the first typo! The output offset voltage created is 400mV, not 200mV as I wrote above. Below is a closeup of the peaks of channel 1 (with the DC servo) and channel 2 (without the servo) that shows it the 400mV difference.

The data sheet for the NJM4556 shows a typical 180nA input bias current

NJM4556 data sheet

which would be 360nA for the two of them in parallel on each channel. That would give

(40.2k)(360nA) = 14.4mV output offset. So the NJM4556 chips are doing better than their "typical" numbers at 3mV for the chips I've measured. For the LME49720 chip

LME49720 data sheet

the typical number is 10nA for one, or 20nA for the pair giving

(40.2K)(20nA) = 0.8mV offset voltage.

I also wanted to point out one interesting thing about the servo, in general, in this circuit. Usually with DC coupled amplifiers the DC servo feedback is applied at the amplifier input, then ripples through all the amplifier stages to the output.

But in the case of the O2 I was able to apply the DC servo feedback to only the unity-gain output stage thanks to the C13 and C14 coupling capacitors that RocketScientist has to DC isolate the two stages. This means that the DC servo would not correct for any DC offset errors in the O2 first stage cause by someone inputting DC from their source. The first stage output would still DC-shift, causing clipping. So the design goal of this servo circuit is mainly headphone protection, not signal correction for applied DC from the source.

Another thing to point out is that the DC servo does effectively become a virtual ground for the 40.2K resistor in this DC servo configuration. That should cause less trouble in this case due to the relatively small current flowing through it and the unity (1) gain of the stage. The main purpose of the 40.2K resistor is to provide a DC return path for the op amp input bias current. Ideally it would be smaller to further decrease the resultant offset voltage created, but it can't be since the resistor forms a parasitic highpass filter with the C13 and C14 coupling capacitors.
 

Attachments

  • O2 amp DC servo ILOAD=10uA on Ch1 closeup.png
    O2 amp DC servo ILOAD=10uA on Ch1 closeup.png
    19.9 KB · Views: 95
Last edited:
O2 bass boost modification

jschristian44 just posted about adding bass boost to the O2. Here is a modification to mess around with that should do the job.

The circuit is essentially the same as figure 15 in the op amp headphone amplifier article on the old Headwize forum, here:

HeadWize - Project: Designing an Opamp Headphone Amplifier (A HeadWize Design Series Paper)

The idea is to add a few parts to the feedback loop on the gain stage op amp to make the voltage gain increase at low frequencies. In Figure 15 if the "bass boost" switch is closed (bass boost off) it shorts out capacitor C3. That just leaves the two resistors R3 and R2 (15K and 56K) in parallel, as the feedback resistor that sets the gain for that non-inverting op amp stage. The parallel resistors are in turn in parallel with compensation capacitor C2. That reduced circuit looks exactly what we have in the O2 gain stage, the 1.5K feedback resistor in parallel with the 220pF comp cap. C1 in the Headwize Figure 15 isn't used here. That capacitor adds very low frequency gain cutoff (in addition to the existing high frequency gain cutoff) to make a bandpass stage, something not needed here.

But when the bass boost switch is opened (bass boost on) in Figure 15 a capacitor is now in series with R3. At low(er) frequencies that capacitor will start to look like an open circuit (increase in impedance), essentially removing R3 from the circuit. That leaves the larger R2 as the only feedback resistor, which in turn increases the stage gain since a bigger feedback resistor gives a bigger gain by the non-inverting gain formula (1 + Rf/R1). The net result is increased stage gain at low frequencies - which is bass boost. :)

So in the case of the O2 the first schematic below shows the parts added to the first (gain) stage of one channel. The second schematic is just a close-up of the bass boost circuit addition. The feedback resistors R16 and R22 are replaced by 7.5K resistors. That resistor in turn has a 2K resistor in series with a 0.1uF capacitor soldered right over it, in parallel. Then two wires would solder across the 0.1uF capacitor can go out to one half of a panel mounted DPST switch that would short out that 0.1uF when the switch is closed for bass boost "off". Opening the DPST switch to leave the 0.1uF caps in the circuit would be bass boost "on". Note that with the boost switch off, at audio frequencies high enough that the impedance of the 0.1uF is much less than 2K, the parallel combination of the 2K and 7.5K resistors give 1.58K, reducing back to (nearly) the original O2 first stage circuit of the 1.5K feedback resistor in parallel with the 220pF compensation capacitor.

The first plot below is the gain through both stages of the O2 with the standard parts values. The second plot is with the bass boost. Compare that plot to the one shown in the Headwize article. You can see that the boost is all below 1kHz in both with the 3dB point at about 200Hz. The bass boost gain is about 10dB. In the case of the O2 that is 10dB added to the existing stage gain. R17 and R21 are 1K in this simulation, so the basic stage gain is 2.5x. Adding the boost circuit causes significant phase change over the frequency sweep as shown, something that wasn't shown in the Headwize diagram.

Note that I've only shown the boost modification on one channel here. It would have to be installed on both channels of course, with one half of the panel mount DPST shorting out the 0.1uF capacitor added to each channel.
 

Attachments

  • AC plot with bass boost.png
    AC plot with bass boost.png
    23 KB · Views: 181
  • AC plot standard circuit.png
    AC plot standard circuit.png
    20.1 KB · Views: 239
  • Circuit with bass boost closeup.png
    Circuit with bass boost closeup.png
    24.7 KB · Views: 274
  • Circuit with bass boost.png
    Circuit with bass boost.png
    40 KB · Views: 272
Last edited:
O2 bass boost mod update

A couple of updates on the bass boost mod, above:

1. A 6.2K resistor would be even better to use for the feedback resistor than 7.5K. Results in nearly an exact 1.5K parallel combination. The closest 1% low noise resistor the right length at Mouser is a 6.19K. So here are links for the 6.19K, 2.0K, and 0.1uF parts at Mouser:

RN50C6191FB14 Vishay/Dale Metal Film Resistors - Through Hole

SFR16S0002001FR500 Vishay/BC Components Metal Film Resistors - Through Hole

BFC241641004 Vishay/BC Components Polypropylene Film Capacitors

2. 3.01K resistors can be used in both positions to produce a smaller 3dB amount of bass boost, also yielding the 1.5K in parallel.

SFR16S0003011FR500 Vishay/BC Components Metal Film Resistors - Through Hole
 
Not sure where else to put this

But what else do you do when your hobbies include, Bow hunting, knife making, electronics and leatherwork.

This pic shows the O2 in a leather case, a small amp in leather and my colorfly likewise in a leather belt clipped case. The O2 case uses the Bamboo front panel.

141518905.jpg


This is just the O2

141518921.jpg


And one more for kicks.

141518922.jpg


I usually use the tactical clip on them for wearing on the belt, I just got to figure out how to keep the amp and source together....I see another man bag that holds my iPad, the amp and the source. Might as well add a 5amp Li-Po battery pack.

I think I am sick,very very sick.

I have a nano, and I am now building a transportable stereo system that will weigh 2kgs. I need an intervention.

Tony
 
Yeah it is a giggle, the bamboo and leather work well together and look even better in the flesh.

I got the idea from when I was into HAM radio and we had "transportable" radios. These were higher power VHF rigs for mountain topping etc. I built one into a leather case together with power supplies and antenna tuners etc. To save weight I de-boxed all the equipment and built separate compartments in a satchel. Still weighed a lot, but it had fan slots etc.

No wonder I am a steam punk tragic as well.
 
I used to have a ham radio license too way back when. A two meter + 440mHz handheld Kenwood with autopatch was my baby. :) Coolest thing on the planet back then to have autopatch access, 7+ years before cell phones, lol.:D At one time I had one of the original TAPR TNC packet data boards set up too.

Hey that is great that you built a portable high power VHF rig! Talk about carting around some really big batteries.
 
Last edited:
Yeah, you know cell phones were what killed radio for me. When you make an international call from a small cell phone the whole radio DX stuff meant nothing.

We did moon bounce for Magellan's centenary and everyone was so disinterested. so you bounced a VHF signal off the moon and used Morse code between AU and the UK? I can have a video call bah bah bah.

The signal was like 30 db below the voice floor, only CW could have done it. I was into sat coms and VHF DX in a big way. Never into AF much though, until know.

The magic just went up like in smoke, just like the money that hobby soaked up. I still have an antenna rotator / elevator in the shed that I just can't part with.
 
Roulston: I had a friend back then who was into moon bounce. :) Amazing how well the moon works as a reflector. I saw an old newpaper article from WWII days where the military did some of the first moon bounce stuff from a ship in Hawaii over to the states. That was before satelites, so it was apparently a big top-secret thing for a while. :)

I had some pretty early access to the internet back in the college days, long before it went public-access, which is what killed off the Ham stuff for me. The fun went out of a radio link to Japan when I could just sent an email, lol. Then cell phones!

jschristian44: I'll try to get some photos and a step-by-step writeup on that bass boost up this evening.
 
Last edited:
O2 bass boost modification step-by-step

By request, here is a step by step guide to building the bass boost modification.

This build is for the 3dB of boost as described above, using: two 3.0K 1% 1/8W resistors; 1 0.1uF 50V X7R MLCC ceramic axial capacitor; and 1 miniature DPDT switch. These parts are shown in the first photo for ID.

The second photo shows the leads of one of the 3.0K resistors bent at right angles, with a small 1mm or so space between the bend and the resistor body on each end. The other 3.0K resistor and the 0.1uF capacitor are hooked together by first bending one lead of each into a "U" shape with pliers. Then the two U's are coupled together and soldered at a right angle, as shown, to form a "V" shape. Two sections of wire are then cut, stripped, and one end of each soldered to half of the DPDT switch.

The 3rd photo shows the V shaped resistor and capacitor soldered onto the ends of the 3.0K resistor. The free ends of the "V" assembly are bent in a "U" shape with pliers, hooked around the ends of the 3.0K resistor between the right angle bends and the body, then soldered.

The 4th photo shows a closeup of the switch connections.

The 5th and 6th photos show the two free ends of the wires coming from the switch soldered to each end of the capacitor on the "V" shaped assembly, then the whole thing inserted into the O2 PCB holes for R22 and soldered. Twisting the wires at a rate of 3 turns per inch or so would be a good idea before final mounting of the switch.

The other channel would be done the same way with two more 3.0K resistors and another 0.1uF capacitor soldered into a V with wires coming off the capacitor. The two switch contacts just below the two that were used here (same side of the switch) would go to the wires on the other channel. That assembly goes into the hole for R16.

And the final result is... IT ROCKS! :D I have to say I really like the 3dB bass boost results. Not too much and not too little. Just enough to correct for the natural bass drop-off of the headphones (and probably of my ears). For someone who really is into bass 6dB would be worth trying.

On one hand a person could say that a mod like this goes completely counter to the design philosophy RocketScientist has for the O2, of passing the signal through with no alteration. But on the other hand a case can certainly be made that at some point along the signal chain - source, amp, or separate EQ box - some EQ may be preferable as per personal taste and/or correction of component issues like bass dropoff in a certain model of headphone.

Enjoy! :)
 

Attachments

  • IMG_1375.JPG
    IMG_1375.JPG
    129.8 KB · Views: 666
  • IMG_1376.JPG
    IMG_1376.JPG
    135.2 KB · Views: 632
  • IMG_1377.JPG
    IMG_1377.JPG
    124.2 KB · Views: 257
  • IMG_1382.JPG
    IMG_1382.JPG
    232.6 KB · Views: 314
  • IMG_1381.JPG
    IMG_1381.JPG
    186.9 KB · Views: 330
  • IMG_1378.JPG
    IMG_1378.JPG
    175.1 KB · Views: 338
  • IMG_1383.JPG
    IMG_1383.JPG
    175.9 KB · Views: 291
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.