Anyone else tinker with overclocking these?

I have 2 so far and 2 more on the way tomorrow, all used. The 2 I already have are newer ones (april & june 2017) with Orange Pi controllers, running firmware v1.1.1.  Using the login creds previously mentioned in this thread (root/innosilicon), I am launching the cgminer manually with a clock value of 1600 instead of 1248.  This results in the hashing rate being around 364 instead of 276.  I kept the voltage at the default value of 820.  So far stable, temps are virtually unchanged.  Can't get any higher than 1600 tho, I think that might be because the PLL or something runs at 1600 as well (saw something about this in the log).  I figure if my temps are normal and voltage was not bumped any higher, how bad could it be. Result is about a 32% overclock.  I'm assuming this is only possible because I have newer batches, so don't take this as a good reference point if you have older ones.  Just wondering if anyone else was tinkering with overclocking these to squeeze every last ounce out of them, and if so what you did to get above 1600.

Another thing I'd like to try is get ahold of the firmware from an A4+ once they're around - wondering if the newer build of cgminer in there might have some improvements that could be copied over to this build.

Go to the web page, stop mining if you have it already mining.  Putty into the miner, change directory to /var/www, and here is the command I use (changed the worker name in case anyone wants to donate some mining time to me in return for this info, so if not change that part to your own pool info)

sudo ./cgminer -o stratum+tcp://scrypt.usa.nicehash.com:3333/#xnsub -u 3Qk7dwXXn1Qr2jfK2aetLZod43v5sJy4ow.Donate -p x --A1Pll1 1600 --A1Pll2 1600 --A1Pll3 1600 --A1Pll4 1600 --A1Pll5 1600 --A1Pll6 1600 --A1Vol 820  --diff 8 --api-listen --api-network --cs 8 --stmcu 0 --hwreset > cgminer.log  &

Hey,

Recently I bought Innosilicon A4 Dominator miners but I'm really unsatisfied with my Nicehash results. Miners are updated with suggested firmware but still I'm getting very low real hashrate. This guy here is getting stable 272MH/s but im not even close to this result... I don't know from what batch are my miners.

Miners are working at 825MH/s total power but nicehash medium result from 6 hours is only 696MH/s - graph from nicehash. Im using xnsub parameter. Here is my miner statistics page from one node - I have two of them.
Because of controller failure I have 3 tubes in one controller but with two tubes in one controller results are that same weak... My controllers are TF card - Orange Pi. Ping to nicehash is 80ms.
Can you show me your results or have some suggestions to me?

I'm really disappointed, was wondering to invest more for L3+ but calculations resulted in favor of A4 products but not in this case. L3+ users are getting stable 488MH/s out of box when miner speed is 504MH/s!

Here are some cgminer summary:
[2017-10-10 23:22:16.876941] BA20 | (5s):68.92M | A:1998848 R:65536 HW:1 WU:48482.6/m
[2017-10-10 23:22:16.877127] BA21 | (5s):68.92M | A:2490368 R:229376 HW:78 WU:63873.8/m
[2017-10-10 23:22:16.877305] BA22 | (5s):68.92M | A:2228221 R:65536 HW:41 WU:53869.4/m
[2017-10-10 23:22:16.877495] BA23 | (5s):68.86M | A:1802239 R:32768 HW:71 WU:43095.6/m

[2017-10-11 00:27:11.848768] BA21 | (5s):70.30M | A:12025855 R:1048576 HW:136 WU:58184.4/m
[2017-10-11 00:27:11.848958] BA22 | (5s):70.30M | A:12877824 R:786432 HW:77 WU:60809.3/m
[2017-10-11 00:27:11.849159] BA23 | (5s):70.30M | A:10190848 R:393216 HW:1019 WU:47101.7/m
[2017-10-11 00:27:11.849356] BA24 | (5s):70.30M | A:11829247 R:589824 HW:705 WU:55267.9/m
[2017-10-11 00:27:11.849541] BA25 | (5s):70.30M | A:12353536 R:294912 HW:743 WU:56872.0/m

Really thanks for any help.