This is free study of HID ballast mainly Mazda3's OEM Mitsubishi G4,5,6 ballast.
These days, it's not rare to hear about over 50W, 55W high wattage ballast, and standard 35W ballast (aftermarket) are sold all over online market in suspiciously cheap price...
It seems so easy to get unreliable quality ballast if you don't know about it or place to buy. (I will mention about these ballasts at later section)
I got little sample data of few ballasts I could obtain.
First of all, I wanted to see OEM ballast performance.
Mitsubish G5 Ballast (Cold Start)
50mΩshunt resister inline
![]()
Amazingly peak of initial boost period's max current is strictly controlled at around 11A. This is because that Mitsubishi ballast has max current control circuit independently and working proper to protect ballast and bulbs.
Once it warmed up enough and stable, 3.4A was measured in average. Assume input voltage 12V, 12V x 3.4A=40.8W
According to ballast data, output supposed to be 34 to 35W its efficiency is about 86%.
(OEM ballast is actually consuming 40 to 42W and outputting 35W)
Mitsubishi G5 Ballast ( Hot Start)
50mΩshunt resister inline
![]()
Since burner is warmed up enough, initial boost control is much less than cold start time.
Here are burner current and voltage data of Mitsubishi G5 Ballast
Mitsubishi G5 Ballast (Cold Start)
50mΩshunt resister inline
![]()
*initial spike is charging current of condenser in ballast and it's not related to ballast performance
At ignition moment, (it is practically open circuit) extreme high voltage was measured.
Soon after light up, voltage goes down to around 30V, and then coming back to 85V, which is stated on ballast.
Right after light up, 2.5A of current was measured, and then it's going down to about 0.41A
At stable condition, 85V × 0.41A ＝ 35W in average.
Right after light up, 30V × 2.5A ＝ 75W (include loss at inverter)
Power is boosted by initial boost control.
Mitsubishi G5 Ballast ( Hot Start)
50mΩshunt resister inline
![]()
For reference,
Here is another sample data from Koito (Denso G3) ballast
Denso G3 ballast (Cold Start)
50mΩshunt resister inline
![]()
Due to initial boost control, peak of max current reaches up to 14A, and unlike Mitsubishi ballast, there was no individual max current control circuit, above 8A period, it seems little overworking tendency. However, Denso ballast also has its own protection system against any abnormal operation as well. This means even 14A overwork looking spike were measured, all circuits still have enough capacity to handle them no problem.
In short, they just have different protection/system design.
I am not sure how this difference affects the bulb though. I personally think Mitsubishi ballast is very well designed and reliable.
Denso G3 ballast (Hot Start)
50mΩshunt resister inline
![]()
Each make have different tendency, but all of their data shows they are easily be able to pass ECE99 code ( lumen need to reach upto 25% at 1 seconds, 80% at 4 seconds of normal operation lumen)
Now I would like to extend study little more to the cheap aftermarket HID ballast.
As I mentioned at beginning, if you are look at online market, there are tons of suspiciously cheap HID conversion ballast like below $50 or so.
Are these really bright, safe, reliable, why so cheap? That’s my simple question.
Sometimes we even see ballast says 50, 55W high wattage HID ballast at also extremely cheap price.
But many of them, just looking at printed labels are already giving me suspicion.
Something like this…
![]()
It says this is 55W HID ballast…
Label says “I norm 3.2A”, I guess this indicates input current of stabilized operation condition. Also (V norm) stated as 13.2V, which means total INPUT performance is
13.2V x 3.2A=42.24W
From this 42.24W, subtract any operation consumption decide ballast out put (burner input),
According to other OEM ballasts analyze, estimated efficiency is about 80% at higher average. I am not sure if I can apply this high efficiency OEM number to those cheap ballast, but let’s say they do have 80% of efficiency, still final ballast out put will be estimated only
42.24W x 0.8(80%)=33.8W
This is like same as 35W OEM ballast actual OUTPUT. Of course it can’t be any brighter than standard 35W OEM ballasts. These are totally misleading information to fake ballast performance seems high. In this way, many OEM 35W ballast should be indicated as 40W ballasts.
Now, I am curious WHAT IS THE FACT going on inside of these cheap ballasts.
Even with cheaper labor, high efficiency factory from China factory, there is no way they can cut actual parts cost radically. Which indicates me there is cost cut at circuit design, omit producing process, QC, use cheaper similar parts for these ballasts to just perform ignite HID bulbs.
If I were getting HID ballast, even if it is USD one, I still get OEM ballast.
Here I got few sample ballasts of 35W and 55W (cheaper ballast from china market), also HELLA ballast to compare performance.
![]()
could it be some sample ballast,,? No labels on them but they look like ballast I showed above.
Opened 55W
![]()
It seems to have minimum caulking but you can tell they tilt ballast case and tried to conserve them. Well, most of high voltage device are on its left side, this kind of make sense… and there are few unsecure bridged wiring too…
Circuit is 2 compositions. (Main circuit and control circuit)
Opened 35W
![]()
This one doesn’t have caulking somehow..
After removing calking of 55W ballast, here is compare picture of 35W and 55W
![]()
Right hand side is 55W one.
Both of them have same parts (resister volumes etc). Even main trans spec is same (normally use different tans for different output rating)
Main switching FET was different but almost all of their specs were same.. I really scared this is high wattage ballast?? Might be they just changing few resisters..
Of course no significant max current control like Mitsubishi ballast.
By the way, HELLA genuine OEM ballast is totally different structure from those two suspicious ballasts. They just look similar on case (cheap one is just mocking casing)
Here is in side of HELLA ballast to compare
![]()
![]()
Back to subject..
Here is control circuit board of 55W one.
![]()
![]()
It seems to have decent IC, this is the one probably main control module. On IC, there is HYLUX letter. Supposedly it is some German company but, unfortunately I couldn’t figure it out. Suspicious..
Anyway, let’s using lumen tester and check these ballast performance.
For reference, HELLA’s OEM ballast set as 100% standard.
(Many cheapwer ballasts are only mocking HELLA’s ballast casing and I never seen some product copy well even inside structure. BTW, Philips ballast is HELLA’s OEM, so they are exactly same)
Result
Ballst HELLA 35W 55W
Brightness 100％ 80％ 134％
Input V 11.89V 12.12V 11.62V
Input A 3.57A 2.98A 4.82A
System Comsumption42.4W 36.1W 56.0W
Ballast OUTPUT 35W 28W 47W
·Test burner: Philips 85126( D2R)
Well, just as I expected, they really don’t have actual output as stated in their label.
There is no point indicating INPUT volume since they are varies depend on ballast efficiency. If ballast only has 64% efficiency, even 55W input end up 35W output. Nothing is deferent from OEM standard 35W ballast.
It’s also important to know these results are from with 1 minutes of ignition. As ballast operate, created heat bring its efficiency down without proper heat control design (not talking about existence of heat sink, but parts layout, parts itself, control circuit)
Another fact from these test also showed that those cheap ballast couldn’t meet ECE99 code (lumen need to reach up to 25% at 1 seconds, 80% at 4 seconds of normal operation lumen). For 55W one, it actually reach up to almost 35% of normal operation lumen in less than second, but right after it drop down lots and took so long to reach 80% of normal operation lumen.
If you are looking to try HID, make sure you get them from reliable source. That’s all I have to say. And OEM ballast does have reason to be OEM. Quality, reliability, protections, they are really on top of those.
These days, it's not rare to hear about over 50W, 55W high wattage ballast, and standard 35W ballast (aftermarket) are sold all over online market in suspiciously cheap price...
It seems so easy to get unreliable quality ballast if you don't know about it or place to buy. (I will mention about these ballasts at later section)
I got little sample data of few ballasts I could obtain.
First of all, I wanted to see OEM ballast performance.
Mitsubish G5 Ballast (Cold Start)
50mΩshunt resister inline
Amazingly peak of initial boost period's max current is strictly controlled at around 11A. This is because that Mitsubishi ballast has max current control circuit independently and working proper to protect ballast and bulbs.
Once it warmed up enough and stable, 3.4A was measured in average. Assume input voltage 12V, 12V x 3.4A=40.8W
According to ballast data, output supposed to be 34 to 35W its efficiency is about 86%.
(OEM ballast is actually consuming 40 to 42W and outputting 35W)
Mitsubishi G5 Ballast ( Hot Start)
50mΩshunt resister inline
Since burner is warmed up enough, initial boost control is much less than cold start time.
Here are burner current and voltage data of Mitsubishi G5 Ballast
Mitsubishi G5 Ballast (Cold Start)
50mΩshunt resister inline
*initial spike is charging current of condenser in ballast and it's not related to ballast performance
At ignition moment, (it is practically open circuit) extreme high voltage was measured.
Soon after light up, voltage goes down to around 30V, and then coming back to 85V, which is stated on ballast.
Right after light up, 2.5A of current was measured, and then it's going down to about 0.41A
At stable condition, 85V × 0.41A ＝ 35W in average.
Right after light up, 30V × 2.5A ＝ 75W (include loss at inverter)
Power is boosted by initial boost control.
Mitsubishi G5 Ballast ( Hot Start)
50mΩshunt resister inline
For reference,
Here is another sample data from Koito (Denso G3) ballast
Denso G3 ballast (Cold Start)
50mΩshunt resister inline
Due to initial boost control, peak of max current reaches up to 14A, and unlike Mitsubishi ballast, there was no individual max current control circuit, above 8A period, it seems little overworking tendency. However, Denso ballast also has its own protection system against any abnormal operation as well. This means even 14A overwork looking spike were measured, all circuits still have enough capacity to handle them no problem.
In short, they just have different protection/system design.
I am not sure how this difference affects the bulb though. I personally think Mitsubishi ballast is very well designed and reliable.
Denso G3 ballast (Hot Start)
50mΩshunt resister inline
Each make have different tendency, but all of their data shows they are easily be able to pass ECE99 code ( lumen need to reach upto 25% at 1 seconds, 80% at 4 seconds of normal operation lumen)
Now I would like to extend study little more to the cheap aftermarket HID ballast.
As I mentioned at beginning, if you are look at online market, there are tons of suspiciously cheap HID conversion ballast like below $50 or so.
Are these really bright, safe, reliable, why so cheap? That’s my simple question.
Sometimes we even see ballast says 50, 55W high wattage HID ballast at also extremely cheap price.
But many of them, just looking at printed labels are already giving me suspicion.
Something like this…
It says this is 55W HID ballast…
Label says “I norm 3.2A”, I guess this indicates input current of stabilized operation condition. Also (V norm) stated as 13.2V, which means total INPUT performance is
13.2V x 3.2A=42.24W
From this 42.24W, subtract any operation consumption decide ballast out put (burner input),
According to other OEM ballasts analyze, estimated efficiency is about 80% at higher average. I am not sure if I can apply this high efficiency OEM number to those cheap ballast, but let’s say they do have 80% of efficiency, still final ballast out put will be estimated only
42.24W x 0.8(80%)=33.8W
This is like same as 35W OEM ballast actual OUTPUT. Of course it can’t be any brighter than standard 35W OEM ballasts. These are totally misleading information to fake ballast performance seems high. In this way, many OEM 35W ballast should be indicated as 40W ballasts.
Now, I am curious WHAT IS THE FACT going on inside of these cheap ballasts.
Even with cheaper labor, high efficiency factory from China factory, there is no way they can cut actual parts cost radically. Which indicates me there is cost cut at circuit design, omit producing process, QC, use cheaper similar parts for these ballasts to just perform ignite HID bulbs.
If I were getting HID ballast, even if it is USD one, I still get OEM ballast.
Here I got few sample ballasts of 35W and 55W (cheaper ballast from china market), also HELLA ballast to compare performance.
could it be some sample ballast,,? No labels on them but they look like ballast I showed above.
Opened 55W
It seems to have minimum caulking but you can tell they tilt ballast case and tried to conserve them. Well, most of high voltage device are on its left side, this kind of make sense… and there are few unsecure bridged wiring too…
Circuit is 2 compositions. (Main circuit and control circuit)
Opened 35W
This one doesn’t have caulking somehow..
After removing calking of 55W ballast, here is compare picture of 35W and 55W
Right hand side is 55W one.
Both of them have same parts (resister volumes etc). Even main trans spec is same (normally use different tans for different output rating)
Main switching FET was different but almost all of their specs were same.. I really scared this is high wattage ballast?? Might be they just changing few resisters..
Of course no significant max current control like Mitsubishi ballast.
By the way, HELLA genuine OEM ballast is totally different structure from those two suspicious ballasts. They just look similar on case (cheap one is just mocking casing)
Here is in side of HELLA ballast to compare
Back to subject..
Here is control circuit board of 55W one.
It seems to have decent IC, this is the one probably main control module. On IC, there is HYLUX letter. Supposedly it is some German company but, unfortunately I couldn’t figure it out. Suspicious..
Anyway, let’s using lumen tester and check these ballast performance.
For reference, HELLA’s OEM ballast set as 100% standard.
(Many cheapwer ballasts are only mocking HELLA’s ballast casing and I never seen some product copy well even inside structure. BTW, Philips ballast is HELLA’s OEM, so they are exactly same)
Result
Ballst HELLA 35W 55W
Brightness 100％ 80％ 134％
Input V 11.89V 12.12V 11.62V
Input A 3.57A 2.98A 4.82A
System Comsumption42.4W 36.1W 56.0W
Ballast OUTPUT 35W 28W 47W
·Test burner: Philips 85126( D2R)
Well, just as I expected, they really don’t have actual output as stated in their label.
There is no point indicating INPUT volume since they are varies depend on ballast efficiency. If ballast only has 64% efficiency, even 55W input end up 35W output. Nothing is deferent from OEM standard 35W ballast.
It’s also important to know these results are from with 1 minutes of ignition. As ballast operate, created heat bring its efficiency down without proper heat control design (not talking about existence of heat sink, but parts layout, parts itself, control circuit)
Another fact from these test also showed that those cheap ballast couldn’t meet ECE99 code (lumen need to reach up to 25% at 1 seconds, 80% at 4 seconds of normal operation lumen). For 55W one, it actually reach up to almost 35% of normal operation lumen in less than second, but right after it drop down lots and took so long to reach 80% of normal operation lumen.
If you are looking to try HID, make sure you get them from reliable source. That’s all I have to say. And OEM ballast does have reason to be OEM. Quality, reliability, protections, they are really on top of those.
