Backfires can be caused by a number of problems but most of them boil down to the same end cause, too much nitrous reaches the engine before adequate fuel gets there. Most of the causes of this problem are due to the manufacturers not either knowing or caring about the consequences of the components they sell in their kits.
They include;
1) Bad component choice (eg. the use of braided hose)
2) Bad kit design (eg. the length of braided hose used)
3) Component failure (eg. a fuel solenoid failure)
But it can also be caused by operator error.
Dealing with the choice of components first;
1) All nitrous supply pipe is ONLY best suited to power levels close to the maximum that the pipe can flow. The bigger the differential between how much the pipe can flow and how much the metering jet/s can flow the greater the risk of a backfire.
This problem usually shows up when using a dyno. Without a purge, the first few shots (exact number depends on jet sizes and run time) the engine runs rich and makes little to no power, so people assume that's how all runs will be and make mixture adjustments to lean it out or crank up the power. However, eventually the gas in the pipe is replaced with liquid (equals more nitrous) and then the combination of the hard hit caused by US kits and the now leaner mix, results in a backfire.
2) Metering jet location is vital for a number of reasons, one of which is preventing a backfire. When a pair of solenoids are opened in a US kit, a HUGE amount of nitrous fills the pipework leading to the nozzles (and metering jets) very quickly but the fuel takes a great deal longer to do the same thing. This results in nitrous reaching the engine long before the fuel, causing the engine to lean out and backfire.
Now you can get away with this situation on a small dose of nitrous but as you go higher and higher, the lean hit gets bigger and worse. Having jets located at the solenoid outlets substantially reduce the flow of nitrous and therefore the risk of a backfire.
3) Supply pipes between solenoids and nozzles also contribute to the problem. The longer and bigger volume they are the slower the fuel pipe fills up and although the same applies to the nitrous, the net result is a BIGGER DIFFERENTIAL in the time it takes between both reaching the engine and that makes the problem worse.
Moving on to the bad kit design;
1) To maximise convenience and profit ALL other companies keep as many components as possible common to all their kits. They use the same bore and same length nitrous supply pipe, regardless of the power the customer wants and the type of car. They use the same length solenoid to nozzle pipes regardless of power and type of car. Now that wouldn't be so bad if they were the best for the majority of application but they're NOT even the best components for ANY job and they are too long for ALL applications.
2) Relying on one pair of solenoids for a 4 cylinder and 2 sets for an 8 cylinder engine results in long pipes with a large volume (plus usually uneven lengths with other kits) but by increasing the number of solenoids to at least a pair feeding a pair of cylinders, allows the pipe lengths to be minimised.
Now how many companies do you know of that offer such a system design??? lol
The final contribution to backfires caused by Nitrous companies is component failure;
1) If a fuel solenoid fails to open then you have the ultimate lean out, resulting in not only manifold destruction but usually engine destruction. Only Pulsoids are put through an intensive bench test program prior to dispatch, as all other solenoids are mass produced and "assumed" to be OK. The test program we put our Pulsoids through would destroy 99% of other solenoids in a small fraction of the time, yet Pulsoids are in better condition after the test than they were before, because they are effectively "run in".
2) Fuel solenoids can fail to open due to bad build quality, bad coil, unsuitable plunger seat seal material expanding.
Anyone with any sense can see from the above details that ONLY WON systems are DESIGNED to minimise the chances of an engine backfire and as a consequence our customers never suffer from a product induced backfire.
However there is one other factor that can cause even a well design and correctly fitted system to experience a backfire and that's operator error;
1) Operator error becomes a bigger concern the more power you add with nitrous. If you activate the system too early in the rpm range (under unnatural loads), on a small dose of nitrous, you'll get away with it BUT if you do the same thing with a big dose, then you can kiss your induction system and possibly your engine goodbye.
2) Nitrous combustion does not 'like' being restrained, it 'likes' the engine to be free to rev easily, so the pistons can get away from the combustion force and so the combustion gases can get out of the chamber in to the exhaust quickly enough. If the engine is under unnatural loads (like on a brake dyno or climbing a hill in top gear), then the pistons can't move away from the combustion forces quickly enough and the exhaust flow isn't adequate enough to get all the waste gases out of the chamber, before the inlet valve opens again and BOOM you have your backfire as the fresh inlet charge is ignited by the burning exhaust gases.
One perfect scenario that often occurs on a dyno is as follows and should be avoided at all costs;
1) Operator exceeds the desired start rpm
2) Operator lets the engine slow down to correct start point and immediately activates full throttle.
3) Instant BOOM!!!!
The issue here is that the engine components are
decelerating and will be even more reluctant to accelerate away from the increased force caused by the nitrous. All nitrous dyno runs should be started WITHOUT NITROUS (unless you have a progressive system with a low start power) and the nitrous system should ONLY be activated once the engine has OBVIOUSLY started to accelerate.
It would also be wise to try a short test run from mid rpm to ensure the system activates correctly and the engine is willing to accept the nitrous, because although this will not ensure the run will be OK from lower rpm, it will help avoid any obvious issues that could cause damage.
I hope that's answered your question.