Old stock Lynx vs New McWill Screen Recom Power Reg Power Consumption

[Orange-Phantom]

Hello,

 

I thought I'd post up my findings on my McWill Modified Lynx II which has been recapped and has the Recom switching regulation in place vs my Old Stock Lynx II.

 

New Lynx Old Lynx

Volts mAmps Volts mAmps

 

10 311 10 202

9 351 9 221

7.9 397 7.9 252

7.2 446 7.2 280

6.7 466 a 6.7 297

6.4 460 b 6.4 318

6.2 415 c 6.2 328

5.9 344

5.5 348

4.5 316 d

3.8 298 e

 

Notes

 

a. Led is lit solid (modded Lynx II)

b. Led is starting to blink (modded Lynx II)

c. Lowest voltage Lynx will operate without switching off (still playable (just)) (modded Lynx II)

d. Lowest voltage that it is still playable (just) (unmodded Lynx II)

e. Lowest voltage will work (screen to dark to see) (unmodded Lynx II).

 

All voltages above were supplied via the input jack, therefore the voltage supplied to the circuit board is 0.7 less due to the voltage drop of the inline silicone diode.

 

Standby current of the modded Lynx II is 8.9ma at 7.9v (7.2v actual on the board).

Standby current of the unmodded Lynx to is 0.001ma at 7.9v (7.2v actual on the board). I was struggling to actually measure the current draw with the meter that I have.

 

So basically from my findings the my new modded Lynx II is drawing more power than my old unmodded Lynx II. Not what I was expecting but hey, the McWill screen is quite frankly amazing and a much recommended upgrade IMHO. Being fair, to get an accurate average I would have to repeat this experiment on a couple more Lynx II's to rule out any variances with boards/processors drawing more current etc. (The revision of Lynx II boards are the same on both Lynx II's).

 

What is very clear though is how much better the old Lynx power circuit is on Batteries not only in it's lowest voltage in which it will allow the device to operate (although I don't know what the lowest voltage the McWill screen could operate at) but also the stand by current. I've had sets of Enloop rechargeable batteries run flat after 4ish days when my modded Lynx II has not been switched on and it's down to the way that Recom switching regulator is always powered on. (Yes I know I can take the batteries out when not in use but that's a pain and there has to be a better way).

 

What is also clear by the measurements that I've taken is that the old Lynx II power circuit is a true high efficiency switching regulator circuit. The zener diode provides a voltage reference to the Hex Inverter which in turn high frequency switches the Q12 FET. This also explains why there is a 100uH choke (L15) after this FET. The FET is not just an on off switch. If you remove the zener diode then the hex inverter has no reference and thus stops switching the Q12 FET and you get 9V straight into your Lynx.

 

People in various forums have been stating that the "crappy" Zener diode is performing the actual power regulation in conjunction with the Q8 transistor. If this were the case then these components would have to be dissipating the voltage drop as heat, they run stone cold so they are not. Or put it another way, when people have bypassed the original circuit with a linear power regulator, they have to heat sink the linear power reg to dissipate the heat and yet the original Lynx power circuit requires no heat sinks and runs stone cold! The current draw vs voltage figures above are further proof that the original Lynx power circuit is a true high efficiency switching regulator.

 

So I'm going to buck the popular trend and put my Modded Lynx II power circuit back to stock (well nearly stock). I'm going to replace the Zener diode and I have already replaced the switching FET. I'll have the benefit of a power circuit that doesn't drain my batteries when it's off and allows me to get more life out of a set of batteries too.

 

I'm posting this up as an alternative solution as I feel that changing the regulator instead of fixing it is in many cases a backwards step especially if you use a Linear instead of a Switching Regulator. This is a great little forum with many helpful people and some amazing mod's and work e.g. McWills screen mod and SainT's forthcoming SD Cart etc. Hopefully this will of been of some use to my fellow Lynx enthusiasts, or at least it offers up a different school of thought on the subject.

Edited December 30, 2015 by Orange-Phantom

[7800fan]

If you are going to stick with zener diode regulation, you should consider a safety after the 5v regulation in case something fails to prevent unregulated 9v from getting dumped into the system. A crowbar circuit would be small and simple and when the fuse blows you know something has gone wrong with the regulator. Or you could steal Ray's 5v switch he made for C64 power and install that inside Lynx: http://personalpages.tds.net/~rcarlsen/cbm/c64/SAVER/but it will take a bit more room inside.

[leonk]

Personally, I don't know why anyone will play Atari Lynx, GameGear, Sega NOMAD or TurboExpress on batteries. I game with AC adapter with all these handhelds.

 

That being said, I'm glad someone took the effort to measure the power consumption of modern LCD vs old LCD. If I were to guess why the modern LCD is using MORE power than the old one, I'd say 80% of the power is being consumed by the FPGA (Xilinx 3) and not the switching power regulator or LCD screen. The LCD screen (if I'm not mistaken) is LED backlit which is a heck of a lot more efficient than the florescent tube in the original Lynx. BUT, the LCD screen does need to communicate with the Lynx PCB, and the FPGA is doing that for you. It's practically a CPU running there! (probably more powerful than what's on the Lynx itself!). That's where you're losing all the power.

Edited December 31, 2015 by leonk

[7800fan]

The old LCD used CCFL which is power guzzler by itself. There was an old post where someone measured the power draw after disabling CCFL, Lynx was drawing about 120mA. If you used the backlight button, it turned off both the backlight and the LCD and got to 75mA. So the old LCD by itself was using about 45mA and the CCFL was about 150mA.

 

The new LCD with the LED backlighting is probably less than old LCD but the chip did draw the most converting from original Lynx video to work on new LCD.

[ApolloBoy]

Personally, I don't know why anyone will play Atari Lynx, GameGear, Sega NOMAD or TurboExpress on batteries.

Uhhh so that they can play them anywhere including places where a power outlet isn't available?

[Lynxpro]

Personally, I don't know why anyone will play Atari Lynx, GameGear, Sega NOMAD or TurboExpress on batteries. I game with AC adapter with all these handhelds.

 

That being said, I'm glad someone took the effort to measure the power consumption of modern LCD vs old LCD. If I were to guess why the modern LCD is using MORE power than the old one, I'd say 80% of the power is being consumed by the FPGA (Xilinx 3) and not the switching power regulator or LCD screen. The LCD screen (if I'm not mistaken) is LED backlit which is a heck of a lot more efficient than the florescent tube in the original Lynx. BUT, the LCD screen does need to communicate with the Lynx PCB, and the FPGA is doing that for you. It's practically a CPU running there! (probably more powerful than what's on the Lynx itself!). That's where you're losing all the power.

 

 

I thought it would be common sense to conclude it's the FPGA because I thought a few people had modded their Lynxes - long before the McWill screen upgrade - by swapping out the existing LCD's bulb with a modern LED and went from just a few hours of gaming on batteries to 4-8 hours usage….

 

Is there a more energy efficient version of that FPGA?

[leonk]

Are people still taking retro consoles with them to places where there is no power outlet? (Camping, the beach)? For those places, I bring my 3DS or PSP. Easy to emulate retro consoles on them. Lower risk of loosing a $400 console too (after cost of console, new LCD and mod cost that's what it ends up costing to most people vs cheap $40 PSP)

 

To each their own I guess. I personally love restoring these old consoles (even to stock) just to keep them out of landfill. Can't imagine helping the cause of I'm dumping dozens of AA batteries every week.

[cowdog360]

I've done a few of the 5V regulator mods, and I also don't like the effects. Its great to have a modern regulator in there, but each lynx is a little different. Some think the batteries are low (maybe an issue with the noisy converter), and as you said the battery drain issue and inability to run on lower voltages.

 

I'm thinking instead of just building a crowbar circuit for overvoltage protection and putting that in line.

 

Also - What did you use for a zener replacement? I'd like to make that standard on all my lynx refurbs.

 

Here's a great crowbar circuit for reference: http://www.allaboutcircuits.com/projects/c-biscuit-power-crowbar-protection-circuit-for-5v-regulator/

 

I'll probably use that same design and make a nice small package to put inside the lynx. Then we have fused protection against failure!

[Clint Thompson]

If only they made rechargeable batteries, the world would be a better place.

 

[cowdog360]

Yeah, the whole rechargeable thing isn't bad.. but it would be nice to be able to use the original circuitry as it does have less draw.

 

Looking at the original circuit, it sure looks like a 5V shunt regulator to me.

[7800fan]

It's just a transistor and a zener diode in emitter-follower (I think) and what it does is pull the rail from unregulated to 5v by dumping excess down but if either the part fails it will not block excess voltage from getting through.

 

Cheap design but not a safe design. Putting a 5v regulator that does break open when it fails rather than short is safer but as other discovered, not all Lynx works the same and some complains about low voltage with flashing LED.

 

A better solution would be to put something like 5v lipo and inject it in where 5v begins (be sure to put ground before mosfet so Lynx can shut off!) and use the lipo charger to handle charging from 5v source, and also use the same 5v source to power Lynx. A good quality 5v adapter can be found cheap. (check with O-scope on 5v adapter with dummy 0.5A and 1A load to be sure it doesn't show excess ripple and be sure that 5v stays within 5%)

[Orange-Phantom]

If the transistor and zener was dumping any excess voltage then surely some portion of the circuit would be getting hot, would in not? I'll tell you what I'll do guys, I'll borrow an oscilloscope to prove or disprove my own thoughts on it being a high efficiency switching regulator. The lack of heat and the current vs voltage measurements make me think that it is, plus the zener/transistor output is feeding the hex inverter which in turn feeds the FET, which has a choke after the output which again makes me think that it is.

 

I'm not wanting to start a flame war, that's not my intention at all. I've swapped my original Lynx 2 with the amazing McWill screen to a recom regulator, it's just that I'm not that impressed with the results of the new power reg conversion and I'm sure that with the amount of brains on this nice little forum we could come up with a much better solution (although I quite like the Lipo idea).

 

If the Recom power reg fails then I could still have 9+ volts straight into my Lynx and I have to say that the original power circuit has served me well for 25 plus years (that's not to say it doesn't need a bit of maintenance).

 

I'll also have a word with a friend who is a fully qualified electronics engineer so I'll tap his brains too and see what he says. It can't do any harm.

Edited March 18, 2016 by Orange-Phantom

[cowdog360]

Ideally, what I'd like to do for refurbishing Lynxes is the following: Replace Q8 and the zener as a preventative since the parts are older. Next, figure out how to build a crowbar or overvoltage protection circuit and integrate it with a polyfuse or something like that. Then I can feel like I've upgraded the Lynx, added protection, and not changed the design much.

[7800fan]

If the transistor and zener was dumping any excess voltage then surely some portion of the circuit would be getting hot, would in not? I'll tell you what I'll do guys, I'll borrow an oscilloscope to prove or disprove my own thoughts on it being a high efficiency switching regulator. The lack of heat and the current vs voltage measurements make me think that it is, plus the zener/transistor output is feeding the hex inverter which in turn feeds the FET, which has a choke after the output which again makes me think that it is.

 

I'm not wanting to start a flame war, that's not my intention at all. I've swapped my original Lynx 2 with the amazing McWill screen to a recom regulator, it's just that I'm not that impressed with the results of the new power reg conversion and I'm sure that with the amount of brains on this nice little forum we could come up with a much better solution (although I quite like the Lipo idea).

 

If the Recom power reg fails then I could still have 9+ volts straight into my Lynx and I have to say that the original power circuit has served me well for 25 plus years (that's not to say it doesn't need a bit of maintenance).

 

I'll also have a word with a friend who is a fully qualified electronics engineer so I'll tap his brains too and see what he says. It can't do any harm.

 

Give your friend the schematic of the power supply so he'd have a better idea of how Lynx gets 5v and how to fix or mod it to ensure a failure won't fry the Lynx. With the current design, if the part fails, the whole Lynx gets 9v and will be fried easily.

[cowdog360]

Ideally, what I'd like to do for refurbishing Lynxes is the following: Replace Q8 and the zener as a preventative since the parts are older. Next, figure out how to build a crowbar or overvoltage protection circuit and integrate it with a polyfuse or something like that. Then I can feel like I've upgraded the Lynx, added protection, and not changed the design much.

 

Well, I found a suitable replacement, a 1N5229B zener diode. Stuck it in the Lynx, all seems to work well. So in theory, this diode should last another 20 years. Had to use a traditional diode since SMD wasn't easy to find, also used a fairchild one instead of some unknown chinese brand.

 

[CasetheCorvetteMan]

So if the unmodded LYNX will operate on as low as 4.5v, why cant the power bank just be wired straight up to the battery terminals and go from there?

 

Seems logical to me, the OP says the unmodded unit will run that low, wouldnt this be the safest option?